How to Know When You Are Ready for an Exam: One Test

You are ready when you can score at or above your target on a full-length paper you have never seen, under exam time, closed book, marked against the published scheme. Everything else is a proxy. Do that twice on different papers, because one run is a sample of one.

Readiness is a measurement problem, and the instrument closest to hand is a feeling. You look at a page, it seems familiar, and familiar reads as known. This page replaces that feeling with a defined test, a threshold you write down in advance, and a response to each way of falling short.

Why does feeling prepared tell you so little?

Because the feeling is built out of recognition, and an exam asks for retrieval. Those are different operations. Recognition is what happens when material is placed in front of you and something in you confirms it has been seen before. Retrieval is producing that material with nothing in front of you. Almost everything you do with the notes open trains the first and leaves the second untested, which is how hours of work on a topic can end with you unable to start a question on it.

There is a one-minute version of this. Open a worked solution in a subject you believe you are fine on and read it through. You will agree with every line, because agreeing with a correct line is easy. Now close it and write the first line yourself, from nothing. If the opening move does not come, what you had was a sense that the argument was reasonable, which is a different possession from being able to generate it.

Highlighting and rereading manufacture that sense most cheaply, and the reason those two were examined at all is itself worth knowing. Improving Students’ Learning With Effective Learning Techniques, by Dunlosky, Rawson, Marsh, Nathan and Willingham, states the selection criterion outright: “some techniques (e.g., highlighting and rereading) were selected because students report relying heavily on them, which makes it especially important to examine how well they work”. Ten techniques went through that examination, each one scored on how widely its benefit holds up over different learners, materials and kinds of test. Highlighting and rereading both received a low utility assessment. The pair at the top of the scale were one of each kind, a way of studying and a way of scheduling it: “Practice testing and distributed practice received high utility assessments”.

Those scores are about techniques, and they stop there. The review measures how well a technique works. It does not reach the question of how a feeling of preparedness tracks a real mark, and no honest page can hand you that number. The link between the two is this page’s own reasoning, drawn from the recognition and retrieval split above. Highlighting and rereading both keep the material in front of you, which is the condition recognition needs and the condition retrieval never gets. Read your confidence off either habit and you are reading it off the operation an exam declines to ask for, which is reason enough to retire familiarity as the instrument.

What counts as a real readiness test?

One unseen full-length paper, sat under exam time, closed book, in a single block with no pauses, marked against the published scheme. Every condition there is doing work, and dropping one turns the exercise back into revision that produces a number you cannot trust.

• Unseen. A paper you have worked through before measures how well you remember that paper. Keep two papers unopened for this.
• Full length. Pacing and stamina are part of the score, and half a paper says nothing about the questions you would have reached in hour two.
• Timed, in one sitting. Set the clock to the real duration and start it before you read the first question.
• Closed book, no pauses. Looking up one fact mid-paper converts a retrieval test into a recognition test, which is the exact swap this procedure exists to undo.
• Marked to the published scheme. Mark point by point before reading any worked solutions, and refuse yourself credit the scheme does not list.

The repeat step is what separates a reading from a result. A single paper can flatter you because it happened to ask about the three topics you revised last, or wound you by opening with the question type you have never liked. Two papers a week apart cancel much of that. If the scores disagree wildly, you do not yet have a stable number, and a third paper is cheaper than guessing which one was real.

How do you set the score that counts as ready?

Your threshold is arithmetic. Write down the mark your target grade needed on a recent past series, add a margin, and that figure is the bar a practice paper has to clear. Keep it as a mark, since a mark is something a score can land above or below, while a target grade held in your head leaves the comparison nothing to work with.

Set the margin from evidence you already have rather than from a number somebody invented. Look up the boundary for your target grade across the last few series of your own exam and note how far it moved between them. That spread is the distance the bar itself can travel from one paper to the next, and taking the highest of those boundaries rather than the average is the cheapest insurance on offer.

Then add something for the gap between your practice conditions and the real ones. At home you sit at a familiar desk, you are the person marking, and nothing is at stake. In the hall there is a clock you did not set, a question that reads strangely on the day, and whatever your hands do in the first ten minutes. So a mark that only just clears the boundary on a practice paper is not the same thing as being safely above it. That deserves saying plainly, because the relief of clearing a bar by one mark is where this procedure most often gets abandoned.

The honest reading is therefore a band. Comfortably above the boundary on two different unseen papers is ready. Just over it on one paper is on track and unfinished. Below it is a number that tells you where the work is, which is worth more than the feeling it replaced. If a paper comes back a long way under target with weeks to go, the job becomes triage, and deciding what to cut in a subject you are failing is the decision that follows.

What does each kind of bad result tell you to do next?

Four different failures look almost identical on the front page and mean completely different things. Sort your lost marks into one of them before you plan a single session, because running the wrong response is how a fortnight disappears with the score unchanged.

The first row is the easiest of the four to misread, because a question you ran out of time for and a question you could not do leave the same blank space. Run out of time and the obvious reading is that you did not know enough, which buys another week of content revision, when the paper showed nothing whatsoever about the questions you never reached. If everything you attempted scored well, the missing marks are a clock problem, solved by rehearsing starts and enforcing a per-question limit.

The last row is worth taking seriously rather than treating as reassurance. If two unseen papers cleared your threshold, the evidence says you are prepared while your nervous system disagrees, which is an ordinary state carrying no new information about your revision. Working harder in response has no measured gap left to close, and the hours come out of sleep.

Turning a marked paper into a list of things to fix is a separate job from deciding whether the score cleared your bar, and the error log method for a marked paper is how that second job gets done properly.

When should you stop revising?

Stop adding new material once two unseen papers have cleared your threshold, and stop everything heavy in the final 48 hours. Those are two separate stops, and the second one arrives whether or not you have earned the first.

The last 48 hours are worth less than you will want them to be worth. Two days leaves little room for an untouched topic to become secure, and the attempt is paid for out of sleep. What the time is genuinely good for is keeping within reach what you already have: short retrieval passes over material you know, the opening moves of each question type, the formulae and dates and definitions that start answers, and the logistics of the morning.

Four things are not worth doing in that window. Starting a topic you had decided to drop. Sitting a full paper too late to act on the result. Rereading a chapter end to end. Comparing your revision hours with anybody else. The full paper is the most damaging, because its only two outcomes are a score you cannot act on and a score that frightens you the night before.

If you cleared the threshold early and a fortnight is still left, the job changes from coverage to maintenance, because material you stop retrieving gets harder to retrieve. How quickly memory decays without review covers the shape of that and what spacing does to it. And when this paper is one of several, dividing revision time across several exams is where the calendar problem gets solved, since a subject you are ready for still needs enough contact to stay ready.

Where GeniusPal fits, and where it does not

The useful thing any study tool can do here is hand you a number where you previously had a mood. GeniusPal turns a document you upload into a set of questions. Feed it one chapter of the notes you already have, and what comes back is a score in place of a page you recognise.

Be clear about what that is worth. A free set of 10 multiple-choice questions over one chapter is a cheap, frequent check, and it is not a substitute for the full-length timed paper this page describes. Multiple choice asks you to pick rather than to produce, and one chapter is not a paper. Use a quiz for the small version of the question, which is whether this specific topic is retrievable today. Use the unseen timed paper for the real one, which is whether you are ready for the exam. Nothing in the product tells you that you are ready.

The numbers decide how often you can run it. A free account carries two study-set generations for the whole lifetime of the account: 10 questions per set, multiple choice throughout, two complete runs of any set you make, and one exam-style analysis, also for the lifetime. Student costs $14.99 a month and lifts those to 100 generations, sets holding as many as 30 questions, ten exam-style analyses each month, and the option of taking the same set as flashcards or as recall. Quiz, flashcard and recall are the only three study modes there are, and every plan accepts uploads as large as 10 MB.

Two things cost nothing on any plan and matter more here than generating new sets. The daily review carries 20 questions, drawn from questions you got wrong or have not tried yet, which is the cheap repeated checking the final weeks call for. Redrilling a wrong answer is free on every plan too.

Exam-style analysis runs in three steps, and none of them is a prediction. A past paper is read on its own into a saved style that records the shape of that paper. You check that style yourself, and until you have, it shapes nothing. You then pick it on a later upload so the questions generated from your notes follow it. The effect shows up most in written answers, which need a paid plan, so on a free account the wording of a multiple-choice set moves much less. What it changes is the questions you practise with. It says nothing about what will appear on your exam.

Whatever you use to check individual topics, the readiness decision comes from one number: your mark on an unseen full-length paper under time, against a threshold you wrote down before you sat it. Set the threshold this week, sit the first paper early enough to act on what it says, and the remaining weeks get spent arguing with a score instead of a feeling.