Three weeks before the exam, most people do the same thing: they go back to the start of the course and reread until everything looks familiar. It is comfortable, it fills the hours, and it barely moves the score.
Here is what does move it, in the order that matters.
Get a structured learning path in 30 seconds →
The ranking nobody applies
In 2013, Dunlosky and colleagues reviewed the ten study techniques students use most, pooling hundreds of studies, and ranked them by actual utility.
| Technique | Measured utility |
|---|---|
| Testing yourself without the material | High |
| Spacing sessions out | High |
| Mixing problem types | Moderate |
| Explaining the content to yourself | Moderate |
| Rereading | Low |
| Highlighting | Low |
| Summarizing | Low |
The two techniques at the top are the ones nobody reaches for on their own. The three at the bottom are the ones almost everyone uses.
The reason for that gap is well understood: rereading is pleasant and produces a feeling of mastery, while testing yourself is uncomfortable and rubs your face in what you do not know. Roediger and Karpicke showed it in 2006, students who reread reported more confidence than students who tested themselves, and remembered less a week later.
The three weeks, in order
Day 21 to day 14: diagnose, do not revise
The first week is not for learning, it is for finding out where you stand. Take each chapter and try to explain it out loud with the material closed. Sort into three buckets: solid, shaky, missing.
That diagnosis takes a day and determines the next two weeks. Without it you distribute your time at random, and in practice you give too much of it to what you already know, because that is the pleasant part to revise.
Day 14 to day 7: active recall on what is shaky
This entire week runs on retrieval: question, attempt from memory, check. Never the reverse.
Spread the passes rather than grouping them. The Cepeda meta-analysis, published in 2006, pooled more than 250 experiments on spacing: for the same total time, spreading sessions out beats a single block every time. Three twenty-minute passes across three days beat one solid hour.
Day 7 to day 1: real conditions
The last week runs on mock papers, timed, without notes. This is where the problems revision never reveals show up: time management, misread questions, the parts you know but cannot write fast enough.
Mark yourself harshly. A mock paper you grade generously is worthless.
Cramming: what it actually delivers
This deserves precision, because the usual advice is dishonest in both directions.
Cramming works in the very short term. If the exam is tomorrow morning, spending the evening revising will improve your score compared to doing nothing. That is not in dispute.
It collapses afterwards. Kornell compared flashcard learning in spaced and massed conditions in 2009: for identical study time, spacing produces markedly better retention. Crammed material holds for roughly twenty-four hours, then goes.
The practical conclusion: if you are on the eve of the exam, cram, it is the right call at that moment. But never build a semester on it, and know that what you crammed for the November midterm will not be there in January.
Exam day
Three habits are worth several points, independently of how you revised.
Read the whole paper before writing anything. You will distribute your time better and avoid spending twenty minutes on a low-weight question.
Start with what you know. You bank points and get moving without freezing. The order on the paper is not an order of attack.
Keep ten minutes to reread. The most expensive mistakes are misread questions and incomplete answers, not gaps in knowledge.
Revising a subject with no course to revise from
Everything above assumes you know what to revise. Outside school that is often untrue: a competitive exam, a professional certification, a self-taught subject, and you have neither a syllabus nor a plan.
Fastudy covers that step. You enter the topic, an adaptive questionnaire assesses your actual level and the time you have left, then the platform builds a path broken into ordered sections, enriched with the best YouTube videos it finds for each part. The built-in quizzes make you work in recall mode rather than rereading.
You get the list of what has to be covered and the order to cover it in, which is what makes the first week's diagnosis possible.
Frequently Asked Questions
How long before an exam should I start revising?
Three to four weeks for a standard exam, six to eight for a competitive one. The criterion is not total duration but the number of spaced passes you can fit over each chapter. Below two passes separated by several days, you are memorizing for tomorrow.
How do I revise effectively the night before an exam?
Test yourself immediately, without rereading first. List what you cannot produce from memory and give it every remaining hour. It is unpleasant because you start by confronting your gaps, but it is the only way to avoid spending your last hours on what you already have. And sleep: sleep is part of consolidation.
How do I improve my test score when I am starting far behind?
Give up on covering everything. Rank chapters on two criteria, their weight in the exam and your current gap, then handle what is heavy and shaky first. A well-targeted partial preparation beats a complete one you never finish.
How many hours a day can you usefully revise?
Three to five effective hours for most people. Past that, recall quality drops and the extra time mostly produces passive rereading. Two blocks separated by a real break beat one double block.
Should I revise several subjects on the same day?
Yes, two per half-day. Alternating makes the session feel choppier, which is precisely the sign that retrieval work is happening. Spending a whole day on one subject produces a feeling of mastery that does not survive the night.
Are study guides useful at this stage?
Only if they ask questions rather than summarize. Three weeks out, writing guides that rephrase the course eats a lot of time for little benefit. Turn your shaky chapters into lists of questions instead.
Useful resources
- Build a revision path with Fastudy
- Dunlosky, J. et al. (2013). Improving Students' Learning With Effective Learning Techniques. Psychological Science in the Public Interest, 14(1)
- Kornell, N. (2009). Optimising Learning Using Flashcards: Spacing Is More Effective Than Cramming. Applied Cognitive Psychology, 23(9)
- Cepeda, N. J. et al. (2006). Distributed Practice in Verbal Recall Tasks. Psychological Bulletin, 132(3)
- Roediger, H. L. & Karpicke, J. D. (2006). Test-Enhanced Learning. Psychological Science, 17(3)