Quick answer: The revision techniques with real evidence behind them are retrieval practice (testing yourself rather than re-reading), spaced repetition (short sessions spread over weeks, not one binge), interleaving (mixing topics within a session), and worked practice under timed conditions. The techniques students rely on most — re-reading notes, highlighting, copying summaries — feel productive and are among the weakest methods studied. Revision works when it is effortful in the right way: pulling knowledge out of memory, on a schedule, in the format the exam will demand. This guide explains each method, why it works, and how to build a plan around them in the time you actually have.
Revision advice is usually folklore: what a successful student happened to be doing when they succeeded. Cognitive psychology has spent decades testing study methods against each other under controlled conditions, and its conclusions are unusually stable — and unusually at odds with common practice. The methods below are the ones that survive the testing. They share a principle researchers call desirable difficulty: the activities that feel harder during revision (recalling without notes, spacing sessions, mixing topics) are precisely the ones that produce durable memory, because the effort of retrieval is what strengthens it. If your exam is a resit and the stakes feel different this time, this guide still applies — and our resit exam help page covers the preparation structure around a second attempt.
1. Retrieval practice: test yourself, constantly
The strongest finding in the literature: practising recall beats re-exposure by a wide margin. Students who revise by reading, then close the book and write down everything they can remember, retain far more a week later than students who spend the same time re-reading — even though the re-readers feel more confident. Work it in three forms: blank-page recall (study a topic, then reproduce its structure and content unaided, then check and repair); flashcards used honestly (answer before flipping; a card you always recognise but cannot produce is not learned); and past-paper questions, which combine retrieval with the exam’s own format. Apps can automate the scheduling of flashcards — our exam marks calculator will not revise for you, but it will tell you what each exam is worth to your module and degree, which is where planning should start.
2. Spacing: little and often, started early
Memory decays predictably, and each review at the point of near-forgetting flattens the decay. Four 45-minute sessions on a topic across two weeks outperform a single three-hour session the night before on every retention measure — and the gap grows with the delay to the exam. Spacing is the main reason starting early works, and it is a scheduling problem, not a talent: the same total hours, redistributed, buy substantially more memory. Build the schedule backwards from the exam date, giving every topic at least three passes — learn, recall, recall under time — and protect the later passes, which are the ones cramming deletes.
3. Interleaving: mix the topics within a session
Studying one topic solidly before moving to the next (blocking) feels orderly and produces fluent familiarity with each topic in isolation. Mixing problem types within a session (interleaving) feels messier and performs better, especially in subjects where the exam skill is choosing the right method — mathematics, statistics, sciences, law problem questions. The reason is diagnostic: blocked practice tells you which technique to use before you start, so you never practise the decision the exam actually tests. Interleave once basic familiarity exists; block only the very first encounter with a topic.
4. Practise in the exam’s format, under its clock
Knowledge that cannot be produced in the format, at the speed, the exam demands is worth less than it feels. Timed past papers expose the real gaps: the topic you can explain untimed but cannot deploy in 40 minutes, the question type whose structure you have not internalised. Work at least the final third of your revision this way, mark against the scheme where one exists, and treat every timed failure as information about allocation — most students discover their problem is distribution across questions, not knowledge within them. For take-home and open-book formats the same logic holds with different materials: practise finding, selecting and citing under time pressure, since the constraint there is judgement, not memory — our take-home exam help page covers that format specifically.
Two more techniques with evidence: self-explanation and elaboration
The widely cited review by Dunlosky and colleagues (2013), which rated common study strategies for effectiveness, placed practice testing and distributed practice at the top — the methods above — and gave moderate ratings to two more that pair naturally with them. Self-explanation: after solving a problem or reading an explanation, state aloud or in writing why each step follows from the previous one. It converts worked examples from things you recognise into procedures you can regenerate, and it exposes the exact step where your understanding is imitation rather than knowledge. Elaborative interrogation: for factual material, keep asking “why would this be true?” and “how does this connect to what I already know?” — generating the explanation forces the new material into your existing structure instead of leaving it as an isolated line in your notes. Both techniques are retrieval’s close relatives: they work because they make you produce something, and production is what the exam will demand. The underlying picture is the forgetting curve first described by Ebbinghaus: retention falls steeply after learning and each successful, effortful recall flattens the fall — which is why systems like the Leitner box (flashcards sorted into boxes by how reliably you recall them, with failures returning to frequent review) work when they are used honestly.
Building your own questions and flashcards properly
Retrieval practice is only as good as the questions you practise on. Three rules keep self-testing honest. Write questions that demand production, not recognition: “Explain why X causes Y, with an example” retrieves; “Is X the cause of Y?” merely recognises, and recognition is the skill exams least reward. Include the awkward questions — the topics you would rather skip are, by the evidence of your own avoidance, the ones the spacing schedule should surface most often; every revision system drifts towards comfortable material unless you deliberately weight it the other way. Close the feedback loop the same session: an uncorrected error practised three times is learned, so every recall attempt ends with checking against the source and repairing the specific gap, not re-reading the whole topic. For problem subjects, a worked example studied, covered, and reproduced from memory is retrieval practice in its purest form — and the errors in your reproduction show you precisely which step of the method you have not yet internalised.
What to stop doing
Re-reading as a main method — fine as a first pass, near-useless as repetition. Highlighting — it marks importance without processing it; if highlighting helps you choose what to test yourself on later, it has a use, but the highlighter is not the learning. Copying notes out neatly — transcription is not cognition. All-night cramming — sleep loss damages exactly the consolidation revision depends on; the night before an exam, sleep is a revision technique. And studying with solutions open — worked examples are for learning a method, after which the book closes and the method is retrieved, or the fluency you feel is borrowed.
A two-week emergency version
If the exam is close, the principles compress rather than change. Day one: inventory the syllabus, rank topics by exam weight and current weakness, and sit one timed past paper to calibrate honestly. Middle days: spaced retrieval cycles on the highest-weight topics, interleaved, with every session ending in unaided recall. Final days: timed papers only, marked, with error patterns (not topics — patterns: misread questions, unfinished answers, one weak method) fixed in short targeted sessions. No new material in the last 48 hours; protect sleep throughout.
Exam revision questions
How many hours a day should I revise?
Fewer, better hours beat marathon sessions: several focused blocks of 25–50 minutes with real breaks, each block containing retrieval rather than reading, will outperform a distracted ten-hour day. If you cannot recall at the end of a block what you studied at its start, the block was too long or too passive — shorten it and add testing.
Should I revise with music or in a group?
Lyrics compete with verbal material for the same processing, so instrumental or silence is the safer default for reading-heavy work. Group revision helps only when it is structured as mutual testing — quizzing each other from past papers — and harms when it becomes shared re-reading with conversation; the test is whether you leave able to produce more unaided than when you arrived.
Is it too late to use these methods a few days before the exam?
No. Spacing compresses (same day, next day, exam eve) and retrieval works from the first session. What changes at short notice is triage: fewer topics, chosen by exam weight, practised in the exam’s format — as the two-week plan above sets out.
Do these techniques work for essay-based exams?
Yes, with the retrieval unit changed: instead of facts, retrieve arguments — the thesis, the three strongest points, the key evidence and the counter-argument for each predictable question — and practise writing timed essay plans, since planning under time pressure is the skill essay exams actually test first.
When revision alone is not the problem
Sometimes the diagnosis is honest: the material was never understood in the first place, and retrieval practice cannot retrieve what was never encoded. If blank-page recall keeps producing blanks on a topic rather than gaps, you need the topic taught or modelled before you can revise it. That is the point of our exam preparation support — topic triage, model answers to past-paper questions you can study from, and a plan built around your dates, all preparation-side and entirely within your university’s rules. Support is priced on the published scale from £20 per 1,000 words; send your exam dates and module list on WhatsApp for a free, exact quote, and use our assignment deadline planner to stage what remains.