Two weeks before an exam, with more to revise than fourteen days will hold, most students open their notes at page one and read, highlighter in hand. That spends the best days on whatever comes first, and never shows what you would get wrong.
This guide, published by EdProcto, sets out a different fortnight: one day to find out where you stand, eleven of practice tests with the book shut, and two light days at the end. Of the eleven sources cited, we read Okano et al. (2019) and the two web pages in full, and the abstracts of the rest.
With a whole term in hand, use our twelve-week revision timetable. EdProcto's page for students shows the self-tests the plan runs on.
Do practice tests beat rereading and highlighting two weeks before an exam?
Yes. In a review of ten common study techniques, Dunlosky et al. (2013) rated practice testing high, and rereading and highlighting low. Most students use those two, the authors note, yet neither reliably lifts performance.
Timing matters. Roediger and Karpicke (2006) had students study prose passages, then either study them again or take recall tests on them. Tested five minutes later, those who restudied remembered more. Tested two days or a week later, those who had practised recalling remembered substantially more, although restudying had left students more confident. An exam two weeks off is the later test.
What does cramming buy?
The next few minutes, and a feeling of knowing. That five-minute result is the honest case for a last-minute reread: it would work if the exam began as you closed the book.
It buys less than the same hours spread out. Kornell (2009) taught people word pairs by flashcard and found spacing more effective than cramming, which the paper defines as massing study on the last day before the test.
How does spacing work inside two weeks?
The gaps shrink to fit: two to five days between sittings on a topic, not the weeks a term allows.
Cepeda et al. (2008) had more than 1,350 people learn facts, go over them once after a gap, and sit a final test as much as a year afterwards. The longer the wait before the test, the longer the best gap, but as a share of the wait it fell: about 20 to 40% with the test a week away, which is a day and a half to three days, and about 5 to 10% with it a year away. For a wait of ten to thirteen days the abstract gives no figure, so two to four days is our reading. No study we read tested a fourteen-day schedule like the one below: it is ours.
Should the topics be mixed?
Mix the re-tests, and know what the evidence covers. Brunmair and Richter (2019) pooled 59 studies of interleaving, where kinds of item are mixed instead of practised in blocks. The benefit was moderate overall (g = 0.42), small for mathematical tasks (g = 0.34), unclear for expository texts, and reversed for words.
So mix the problem types that are easy to confuse: the exam will not announce a question's topic. Our guide to interleaving goes further.
How much does sleep matter in the last two weeks?
More than the last night alone. Okano et al. (2019) tracked the sleep of 88 students through a university chemistry course. Longer, better and more regular sleep went with higher marks. Sleep on the single night before a test showed no relation to its result, while sleep over the month before each midterm did, as did the week before some quizzes. It is a correlation in one course, and the authors say showing cause would need a randomised trial.
Gillen-O'Neel et al. (2013) collected daily diaries from 535 school students. After a student gave up sleep to study more than usual, they had more trouble understanding lessons the next day and were more likely to struggle on a test or an assignment.
The NHS (2023) says most teenagers need 8 to 10 hours a night and calls cramming all night before an exam usually a bad idea. So every day of the plan ends by your usual bedtime.
Why is a failed practice test useful?
Because trying, failing and then seeing the answer teaches more than seeing the answer alone. Kornell et al. (2009) used questions that could not be answered, some of them about invented events. People who tried first and were then shown the answer learned more than people shown question and answer together.
So a low score on day one is the diagnostic working, provided each miss is corrected the same day: see our guide to the questions you got wrong.
Does practising under exam conditions help with nerves?
The NHS (2023) advises it. It calls nervousness a natural reaction to exams and, where anxiety gets in the way, suggests practising what the day will ask, such as practice papers under exam conditions. Anxiety or low mood that is severe, persists and interferes with everyday life is, it says, a reason to get help, starting with a GP.
That is advice, not a trial. Our guide to past papers covers an experiment on practice tests and stress, and found no controlled study of timed against untimed papers.
How do you decide what to drop and what to secure?
Decide on day one from two documents, the exam's published outline and one diagnostic quiz, not from how well you feel you know a topic.
The outline is the specification, syllabus or test plan the exam's setter publishes. AQA (n.d.), for example, sets out GCSE Biology as eight numbered topics, with Paper 1 covering topics 1 to 4 in 1 hour 45 minutes for 100 marks. List your topics and call each heavy or light: heavy if it carries at least an even share of the marks. Where no weights are printed, count each topic's marks in the last two papers: our working rule.
The diagnostic quiz is about five questions on every topic, answered in one go with the book shut. Score each topic separately, counting a lucky guess as a miss, then sort:
| Diagnostic score | Heavy topic | Light topic |
|---|---|---|
| 4 or 5 of 5 | Secure | Secure |
| 2 or 3 of 5 | Win | Win if a slot is left |
| 0 or 1 of 5 | Win, core only | Drop |
A secure topic is only re-tested. A topic to win is studied, then quizzed, and the plan has four slots for them. "Core" means the definitions and the one or two standard question types.
The table is our working rule, and five questions is a coarse reading. Son and Metcalfe (2000) found that people with too little time to study everything gave more of it to the items they judged easy. That is what learners choose under pressure, not proof that the choice is best. Our reasoning is cost: a half-known topic needs less time for each mark it can still add.
How do you revise for an exam in two weeks, day by day?
With one day of triage, six of first quizzes and early re-tests, five built round two timed papers, and two light days. Day 1 is fourteen days before the exam, which falls on the morning after day 14. W1 to W4 are the topics to win, heaviest first; S is the secure pile.
| Day | What to do |
|---|---|
| 1 | Outline, diagnostic quiz, sorting table. EdProcto writes the diagnostic from your notes |
| 2 | Study W1 for an hour, shut the book, take its first quiz |
| 3 | The same for W2 |
| 4 | The same for W3. Re-test W1 |
| 5 | The same for W4. Re-test W2 |
| 6 | Re-test W3, and W1 if day 4 went badly |
| 7 | Half day: re-test W4 and S (its diagnostic quiz), mixed |
| 8 | Timed paper 1, marked the same day |
| 9 | Repair the two weakest areas the paper found, with a new quiz on them |
| 10 | Mixed re-tests: W1, W2, W3 |
| 11 | Mixed re-tests: W4, S, the repair quiz |
| 12 | Timed paper 2, marked the same day |
| 13 | Last look: every kept quiz once, in two mixed sittings |
| 14 | The last day: misses only |
Three rules:
- Every sitting ends with a retry. Read each correction, then answer the misses again from memory until none is left.
- A re-test is the whole quiz, not only the earlier misses.
- A bad re-test returns sooner. With three or more wrong in ten, take it again two days on.
The plan assumes two to three hours a day; with less, drop a topic before you drop a re-test. A timed paper is a real past paper where you have one, sat on paper at the exam's own pace: on the AQA outline above, 105 minutes for 100 marks is just over a minute a mark.
What should you do on the last day and the morning of the exam?
Little, and nothing new.
- Morning of the last day. Answer from memory every question still on your list of misses.
- Afternoon. Read the outline once more for the paper's length, marks and question types. Pack what the exam allows and check the time and place.
- Evening. Stop early, leave the half hour or so of winding down that the NHS (2023) suggests, and go to bed at your usual time.
- Morning of the exam. Read your list of misses once, with the right answers beside them. Open nothing new, and arrive early.
The routine is ours.
What does one student's fortnight look like in scores?
Like this, in an invented example: the exam, its outline and every score are made up, and no real student is described.
An economics paper has six topics. Day 1's diagnostic, beside the outline's weights:
- Markets and prices, 25%: 4 of 5, secure (S).
- Market failure, 20%: 1 of 5, win the core (W1).
- Costs and firms, 20%: 2 of 5, win (W2).
- Elasticity, 15%: 3 of 5, win (W3).
- Labour markets, 12%: 3 of 5, win (W4).
- Trade, 8%: 0 of 5, drop.
Then, out of ten:
- Days 2 to 5. First quizzes: W1 4, W2 5, W3 6, W4 7. Re-tests: W1 7, W2 8.
- Days 6 and 7. W3 9, W1 9, W4 9, and S 5 of 5.
- Day 8. Timed paper 1: 52%, the last question not reached.
- Day 9. A repair quiz on externalities and cost curves: 6.
- Days 10 and 11. W1 10, W2 9, W3 10, W4 10, the repair quiz 9.
- Day 12. Timed paper 2, a different year: 63%, with four minutes to spare.
- Days 13 and 14. Three misses in 55 questions, retried that day and on the last morning.
Scores that climb on questions already seen are partly memory of those questions, so the two unseen papers are the better reading. Neither forecasts the exam.
How do EdProcto's self-tests fit the fourteen days?
EdProcto is the best tool for a two-week plan built on your own notes, because time is what the plan lacks and it removes the slowest job: Procto AI writes each quiz when you need it. The study screen then puts the right answer beside every miss, and the dashboard names the quizzes that are due.
- The diagnostic. In the quiz maker on the home page, choose "Test yourself", paste your notes on one topic, set "How many" to 5 and press "Write the quiz". One short quiz a topic, taken back to back.
- A first quiz. The same, with "How many" at 10, then "Take it now".
- Each miss corrected. On the study screen, press "Check" to see "Right.", "Partly right." or "Not this time.", then the answer and why.
- The retry. The end page counts your misses by topic and offers a button such as "Retry the 6 you missed".
- The re-tests. The dashboard lists a quiz under "Due for a re-test" two full days after your latest sitting when the last whole run had a miss, and seven days after a clean one. "Take it again" starts the whole quiz. On any other day, "Study" on its row opens it.
- The clock. In the quiz's settings, "Time limit" at "Set" with "Minutes overall" puts a countdown on the screen.
Read each question against your notes before you trust it: Procto AI is asked to write only from the text you paste, and "Edit the quiz" opens the builder to change one. Paste your own notes; a past paper goes in only where its board's terms allow. With no notes to hand, the kits hold ready-made quizzes by subject and level. More in re-tests that come due.
Which plan the fortnight needs
The study screen, the re-test list, ten quizzes a month and 20 questions a month written by Procto AI come with the Free plan. The example uses 80 written questions in eleven quizzes: 30 for the diagnostic, 40 for the topics to win, 10 for the repair quiz. That is the Learner plan: 100 written questions a month, self-tests without limit, and a whole textbook on the Books page, with a map that marks each chapter "known", "getting there" or "go back".
On Free, take the diagnostic on paper from your textbook's own questions, let Procto AI write ten questions each for the two heaviest topics to win, and type the rest yourself. See the pricing page. No plan and no tool promises a grade.
Quick questions
Is two weeks enough to revise for an exam?
Enough to test every topic you keep four or five times. Whether that reaches the mark you want depends on where the diagnostic puts you, and no plan can promise it.
Should I reread my notes or do practice questions?
Practice questions, with rereading kept for the section behind each miss. Dunlosky et al. (2013) rated practice testing high and rereading low.
Can EdProcto make the practice tests for a two-week plan?
Yes. Paste your notes into the quiz maker, choose "Test yourself" and take the quiz on the study screen; the dashboard lists it again when it is due. The page for students shows each step.
Should I stay up late to revise the night before?
No. The NHS (2023) advises against it, and in Gillen-O'Neel et al. (2013) students who gave up sleep for extra study struggled more the next day.
References
- AQA. (n.d.). GCSE Biology 8461: Specification at a glance. Retrieved October 10, 2026, from https://www.aqa.org.uk/subjects/biology/gcse/biology-8461/specification/specification-at-a-glance
- Brunmair, M., & Richter, T. (2019). Similarity matters: A meta-analysis of interleaved learning and its moderators. Psychological Bulletin, 145(11), 1029–1052. https://doi.org/10.1037/bul0000209
- Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & Pashler, H. (2008). Spacing effects in learning: A temporal ridgeline of optimal retention. Psychological Science, 19(11), 1095–1102. https://doi.org/10.1111/j.1467-9280.2008.02209.x
- Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students' learning with effective learning techniques: Promising directions from cognitive and educational psychology. Psychological Science in the Public Interest, 14(1), 4–58. https://doi.org/10.1177/1529100612453266
- Gillen-O'Neel, C., Huynh, V. W., & Fuligni, A. J. (2013). To study or to sleep? The academic costs of extra studying at the expense of sleep. Child Development, 84(1), 133–142. https://doi.org/10.1111/j.1467-8624.2012.01834.x
- Kornell, N. (2009). Optimising learning using flashcards: Spacing is more effective than cramming. Applied Cognitive Psychology, 23(9), 1297–1317. https://doi.org/10.1002/acp.1537
- Kornell, N., Hays, M. J., & Bjork, R. A. (2009). Unsuccessful retrieval attempts enhance subsequent learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35(4), 989–998. https://doi.org/10.1037/a0015729
- NHS. (2023, March 8). Help your child beat exam stress. https://www.nhs.uk/mental-health/children-and-young-adults/advice-for-parents/help-your-child-beat-exam-stress/
- Okano, K., Kaczmarzyk, J. R., Dave, N., Gabrieli, J. D. E., & Grossman, J. C. (2019). Sleep quality, duration, and consistency are associated with better academic performance in college students. npj Science of Learning, 4, Article 16. https://doi.org/10.1038/s41539-019-0055-z
- Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249–255. https://doi.org/10.1111/j.1467-9280.2006.01693.x
- Son, L. K., & Metcalfe, J. (2000). Metacognitive and control strategies in study-time allocation. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26(1), 204–221. https://doi.org/10.1037/0278-7393.26.1.204