Revision · Exam technique
How to Use Past Papers When You Haven't Revised Everything Yet
A past paper can be a map before it's a mock exam. Here's how to use it when you're still learning, without treating every blank answer as a verdict on your intelligence.
You open a past paper, read the first question, and shut it again. You haven't covered that topic. The second question looks familiar, but you wouldn't know how to begin. By question three, the paper feels less like revision and more like a very official document explaining everything you don't know.
The common advice is “just do past papers”. Another version is “don't touch them until you've learned everything”. Both miss something important. A past paper can be used for three completely different jobs: finding out what's expected, practising particular skills, and rehearsing the whole exam.
You don't have to be ready for the third job to get value from the first two.
Quick answer: what to do with a paper today
- Confirm it's for the correct subject, specification or module.
- Spend ten minutes looking at question types, time limits and marks. You're not being tested yet.
- Choose one question from a topic you've at least begun.
- Attempt it with whatever support fits this learning stage. Write an answer before reading the mark scheme.
- Compare your answer with official guidance, name the gap and retry a similar question later.
Save a full timed, closed-book paper for a later stage. If the exam is close, move to realistic timing sooner and ask a teacher what to prioritise.
The three ways of using a past paper
| Mode | When it's useful | What you do | What success looks like |
|---|---|---|---|
| Map | Early, when many topics are unfamiliar | Inspect structure, command words, weighting and coverage | You know the assessment you're preparing for |
| Workshop | While learning a particular topic | Attempt selected questions and study your errors | You can improve the next answer |
| Rehearsal | When you've covered enough and need exam-condition practice | Complete a relevant paper to time, under the actual permitted conditions | You can apply knowledge, manage time and review performance |
These modes aren't official exam-board categories; they're a practical way to stop using the same resource for the wrong job. If you've only just met quadratic functions, a two-hour timed paper may reveal that you haven't learned quadratic functions. You knew that before opening it. A focused question can teach you much more.
At the same time, don't stay forever in the comfortable “workshop” stage. Learning how to solve a question with a textbook open isn't yet proof you can solve one when the book is closed. You need to make that transition deliberately.
Check the paper actually belongs to your exam
Before you download an exciting pile of PDFs, look up your current syllabus, exam board, module code, assessment brief or specification. Check the exam year, permitted materials, calculator policy, paper length and any changes to the assessment format.
For GCSEs and A levels, start with the relevant examining body's official materials. OCR provides a past-paper finder with question papers, mark schemes and examiners' reports. Its subject adviser explains that some recent papers remain restricted to teaching centres, and a mark scheme has to be read in context. Not finding the newest series publicly doesn't mean you're looking in the wrong place.
University exams vary even more. Your lecturer may provide a specimen paper, problem sheets or a bank of short answers rather than full previous papers. The Open University's guide recommends investigating the actual structure of your module's assessment, including compulsory questions and the number of marks per section. That's a good question for any institution: What exactly am I being assessed on this year?
Older papers can still provide useful topic practice, but don't treat a question from a retired specification as evidence that the same question or weighting is coming back.
Read the paper before trying to finish it
Let's imagine you've found a valid 90-minute paper with three sections: ten short definitions, four calculations and one extended explanation. Don't answer anything yet. Ask what the shape of the assessment is telling you.
It asks you to retrieve precise terms, apply a process under time pressure and explain the meaning of an answer. If your notes only contain definitions, you've discovered a mismatch between how you're revising and what you'll need to do.
Write one simple map: memory, method, interpretation. For an essay-based subject, it might be argument, evidence, judgement. A computer science paper could require recognise, implement, debug. These are examples, not claims about one particular exam board.
Also note the command words. “State”, “calculate”, “compare”, “evaluate” and “justify” ask for different responses. A perfectly memorised fact cannot, by itself, satisfy a question asking you to assess competing explanations.
How to attempt a question you don't fully know
Start with a question related to a topic you've seen, even if you're shaky. Give yourself a short genuine attempt before opening notes. Write the method you'd try, a rough outline, a diagram or the first sentence you can defend.
Then pause and consult the right resource. For a calculation, find a worked example that explains the step you're missing. For a historical comparison, revisit evidence supporting each side. For a biology explanation, confirm the causal sequence rather than learning the mark scheme's words as a little poem.
Close the resource and restart the question on a fresh page. Don't copy the completed solution: the second try is where you discover whether the explanation has become something you can use.
UC San Diego's learning guidance describes retrieval practice as trying to recall before checking the material, then correcting mistakes. This is why the uncomfortable first attempt matters. Reading the answer without trying to produce one tends to feel easier than it really is.
A worked example: a question worth eight marks
The question and marking points below are invented for demonstration. They are not from an official examination, and the marks are illustrative rather than an authentic grading rubric.
Example question: why might two studies on the same topic reach different conclusions? (8 marks)
First attempt: “One study had more participants and one researcher might be biased.” That's a plausible start, but it's not a developed explanation and it doesn't show how the differences affect conclusions.
After reviewing the course material: You notice four issues to consider: how participants were selected, what was measured, how the studies controlled other variables, and how uncertainty affects the conclusion.
Second attempt outline: “Different participant samples may limit how results generalise. Measuring slightly different outcomes can make the studies appear inconsistent even when they aren't testing the same claim. Uncontrolled variables can also create competing explanations. Comparing uncertainty and the study design is necessary before concluding that either study is wrong.”
The next challenge: Check the real module's marking criteria and evidence requirements, then write a complete response within an appropriate time. The improved outline isn't automatically worth eight marks. It's a bridge from vague recognition to reasoning.
Notice what made this useful: the mistake told you what kind of knowledge was missing. It wasn't simply “learn more psychology”. It was “learn how to connect a methodological difference to the conclusion it changes”. That's the size of action your next study block can hold.
Read the mark scheme as feedback, not as a script
Mark schemes are often difficult to read because they are written for marking, not as polished model essays. Some allow equivalent wording; some credit method marks; some offer indicative content rather than one exact sentence. Read the instructions and qualification-specific guidance before assigning yourself a score.
For calculations, distinguish between a wrong final answer caused by an arithmetic slip and a wrong method from the first step. For essays, check whether the assessment values analysis, evidence, relevance or judgement. Highlighting a keyword in the mark scheme doesn't prove the argument earned a mark.
OCR's 2026 guidance suggests predicting the elements required for full marks before examining the mark scheme, then comparing your prediction with the official material. It also explains that examiner reports can reveal patterns that the scheme alone doesn't show.
If you don't have an official mark scheme, use the teacher's rubric, model answers or feedback. For university essays, self-awarded percentages can be especially misleading. Ask for formative feedback on a short plan or sample paragraph rather than assuming an online checklist is the examiner.
The mistake log that doesn't become another subject to revise
A mistake log is only useful if the next session changes because of it. It shouldn't require thirty minutes of colour coding after every ten-minute question.
A small error log you can actually use
| Question | What went wrong | Next repair | Retry |
|---|---|---|---|
| Statistics Q3 | Chose wrong test; didn't identify data type | Practise choosing tests from scenarios | Friday |
| History essay Q2 | Examples listed but not evaluated | Write one evidence-to-judgement paragraph | Sunday |
| Chemistry Q6 | Units converted incorrectly | Three unit-conversion exercises | Thursday |
Illustrative rows. A retry date is a reminder, not a universal scientifically optimal interval.
Try categorising errors as: Didn't know (the idea was missing), Couldn't use (the idea was familiar but wouldn't transfer), Misread (you answered a different question), or Ran out of time (the process was too slow). You can also note careless arithmetic or missing units under “Misread/execute”. These categories point to different remedies.
If a formula is unknown, retrieve and practise it. If you understand a rule but can't spot when to use it, compare several problem types. If you misread “evaluate” as “describe”, practise underlining command words and planning your argument. If timing failed, decide whether it was a knowledge problem, overlong answers or a difficult exam layout before simply trying to write faster.
What to do after a low score
Low marks on a first diagnostic paper can feel very personal. But scores from an untimed, open-book practice question and a closed-book, timed simulation are measuring different things. Don't plot them on the same graph and call the movement “improvement” or “decline”. Label the conditions honestly.
Instead of writing “42%, terrible” at the top, write what the score revealed. For example: “Under timed conditions, I can calculate basic probabilities but lose marks interpreting them.” Now you have a task: write explanations for three solved questions, check them against the rubric, and revisit an unfamiliar variant.
If you see the same error three times, stop doing entire papers for an hour. Investigate the prerequisite, ask a teacher, and practise the exact skill. Finishing more papers while repeatedly making the same mistake can build a large stack of evidence that you need to change methods.
How the method changes by subject
Maths and physics: Show your working even when you can do a step mentally. Many assessments award method credit, and written steps make error diagnosis possible. Mix question types only after you understand the separate methods well enough to identify which one applies.
Biology and chemistry: Check both content and explanation. In sciences, an answer may require a causal chain, units, diagrams or precise distinctions. Memorising mark-scheme phrases without understanding the underlying process can break when the question is worded differently.
Literature, history and essay-heavy subjects: Try five-minute essay plans early. Identify the claim, supporting evidence, counterargument and judgement before attempting full-length essays. A mark scheme may use broad levels or indicative content; it isn't always a shopping list of facts.
Languages: Practise the actual skill assessed. A writing paper is not the same as listening comprehension, and you can't prepare for speaking solely by reading transcripts. Use appropriate audio and speaking exercises, plus any official sample tasks.
University open-book exams: Rehearse finding and applying information within the permitted materials, not just locating a sentence. “Open book” doesn't remove time pressure or the need to understand what the question asks. Check university rules about collaboration and AI-assisted tools; don't invent your own interpretation of what is allowed.
When should you start full timed papers?
When you've covered enough content that the result will tell you something other than “I haven't learnt half the course”, start rehearsing the real conditions. Don't wait for perfect confidence; it often arrives later than is useful. But don't confuse early diagnostic work with an exam rehearsal either.
A practical progression might be: inspect one paper early in the term, attempt targeted questions while learning, then try a full timed paper with enough time left to review and practise weak areas. The exact timetable depends on your assessment date, how many valid papers exist and whether your school or university keeps some for mocks.
Keep at least some suitable questions unseen for a realistic final rehearsal. If you've memorised every available question and answer, a high mark on the same paper may measure recognition rather than your ability to handle a new one. Ask teachers about additional approved practice instead of using restricted or leaked material.
MIT's learning science overview explains why spacing and eventually mixing problem types can improve later learning, even though it may feel harder initially. The goal isn't the smoothest practice session. It's a more dependable answer when the question changes.
A realistic 75-minute past-paper workshop
Minutes 0–10: Verify the source and choose one relevant topic or question set.
Minutes 10–30: Attempt the questions without looking at answers. If this is early learning, pause to consult a worked explanation only after a genuine first try.
Minutes 30–50: Mark carefully using the official scheme or course criteria. Separate content gaps from question-reading and timing mistakes.
Minutes 50–65: Repair the highest-value error with a focused explanation and one related exercise.
Minutes 65–75: Reattempt one question from memory and put a different example aside to retry later.
This schedule is an illustrative study format, not a prescribed duration or a claim about optimal learning time.
Notice that the most valuable part might be the final 25 minutes. If your revision habit is to mark a paper, feel bad or pleased, and immediately start the next paper, you're missing the part where feedback changes your next attempt.
What if your exam is next week and you're not ready?
With little time left, get very clear about assessment coverage and the areas where you can still make progress. Start with a short diagnostic set rather than a soothing read-through of every folder. Choose a mix of fundamental, high-relevance topics and practise the question types that actually appear in your current assessment.
It's tempting to count predicted marks and skip entire areas. That can be risky when papers change, questions are compulsory or topics combine. Use the official specification and your teacher's advice, not an internet prediction of “topics that always come up”.
Don't spend your last night cramming until dawn to rescue one extra paper. A fatigued brain still has to read instructions, plan answers and notice errors. Get realistic rest and food, and protect some time to sort the practicalities: venue, allowed materials, transport and start time.
If you're overwhelmed by the whole syllabus, our broader revision guide can help you plan returns across topics. If you struggle to remember material once the page closes, try retrieval practice beyond rereading.
Where the tools come in, and where they don't
Paper and a pencil may be all you need. A spreadsheet can hold the error log. A calendar can mark your next attempt. There's no need to build a complicated revision dashboard unless it genuinely helps you return to the questions.
In Calphii, which we make, you can keep a small error log in notes, a question to retry in your task board, and a Deep Focus timer beside your current paper. But Calphii won't mark your exam, replace the official scheme or tell you what's coming up. If one printed page is clearer, use the printed page.
Questions people ask before opening their first paper
Is it bad to do a past paper before revising?
No, as long as you know what you're using it for. An early paper can reveal format and priorities. A poor first score isn't evidence that you'll perform equally badly after learning; it is a description of what you could do under those particular conditions.
Should I use notes while attempting questions?
During a learning workshop, yes, after a genuine independent attempt, if it helps you understand a method. During exam rehearsal, follow the real exam rules. Label the conditions so you don't compare unlike attempts.
How many past papers should I do?
Enough to practise the exam's demands, uncover weak points and revisit them. There isn't a responsible universal number. One thoroughly reviewed paper can be more useful than several marked in a rush; later you still need enough varied practice to test transfer and timing.
What if there are no past papers for my course?
Ask your lecturer or teacher for specimen questions, sample answers, assessment criteria and topic-based exercises. You can construct practice prompts from learning outcomes, but don't treat them as predictions of official exam questions.
Can I use AI to mark my essay?
Only in ways your institution allows. A tool can suggest missing points or alternative interpretations, but it can misread rubrics, invent criteria and award convincing-looking grades without an examiner's judgement. Compare against official guidance and seek human feedback for significant assessments.
Leave the paper knowing what to do next
The first time you open a past paper, you may not feel ready. That's fine. You aren't trying to prove readiness to a PDF. You're trying to learn something useful about the exam and about what you can do next.
Circle one error you understand. Repair it. Try again later. Then, as the real exam gets closer, practise answering unfamiliar questions under the conditions that will matter. The paper stops being a verdict and becomes a conversation with the work.