Active Recall vs Rereading: Which One Actually Works?

Quick answer: Active recall beats rereading because retrieving a fact strengthens the memory, while reading it again only strengthens the feeling of familiarity. In the experiment most people quote, students who self-tested remembered roughly 60 percent a week later, against roughly 40 percent for students who reread it — even though the rereaders felt far more confident.

I lost most of Year 10 to rereading. My Cambridge IGCSE history notes were genuinely lovely: three highlighter colours, ruled margins, key dates boxed. I would go through a chapter three times and close the file feeling ready. Then I sat a mock, hit a question on why the League of Nations failed, and produced two thin reasons out of a possible four. History is my favourite subject and I had put the hours in. The problem was never effort. It was the method.

Why rereading feels like it works

Rereading feels effective because it produces fluency, and your brain reads fluency as knowledge. The second time through a page the sentences arrive faster, nothing surprises you, and that smoothness gets misread as "I know this."

But fluency belongs to the page, not to your memory. It vanishes the moment the page does, and in an exam there is no page. There is a blank space and a clock.

Psychologists call this a judgement-of-learning error. Students who reread overpredict their own scores; students who self-test underpredict, feel worse, and score better. That gap is the whole reason rereading survives: it lies to you in the direction you want to be lied to.

Highlighting is the same trick with a pen in your hand. Highlighting a line is one decision about what matters, taking about half a second, and half a second of thought does not build a memory.

What actually happens when you recall

Active recall works because retrieval is itself a learning event, not a measurement of one. Every time you drag something out of your head unaided, you strengthen the route back to it. Two details matter:

• Failed attempts still count. Trying to remember something, failing, then looking it up beats looking it up straight away. The struggle is the mechanism, not wasted time before it.
• Difficulty predicts payoff. The harder a retrieval feels while still succeeding, the more it sticks. Recalling a whole section from one heading beats flipping a card that shows half the answer already.

It also explains why recall feels worse. Rereading hides your gaps; recall puts them on a page in your own handwriting. The discomfort is the useful part, though it took me a fortnight to stop hating it.

The blank page method, worked through with real numbers

Here is the process, with numbers from one of my own sessions on why the League of Nations failed. The textbook section held 14 separate points I could credit myself for.

1. Read the section once, properly. Eight minutes. No highlighter, one pass.
2. Close everything and take a blank page. Write the heading at the top. Five minutes to dump every point you can in note form. No full sentences.
3. Mark it against the source. Tick what you got. Add what you missed in a different colour. First attempt: 6 out of 14.
4. Stop there. The marking is diagnosis, not revision.
5. Next day, blank page again, without rereading first. Four minutes. Second attempt: 11 out of 14, including four of the five I had missed the day before.
6. Day four, one more pass. Three minutes, 14 out of 14.

Total: about 20 minutes over three days. My old method was three rereads at eight minutes each, so 24 minutes, for a next-day blank page that would still have produced roughly 6. The recall route cost less time and roughly doubled the output.

The scoring matters more than it looks. Counting your recall out of a fixed total turns a vague feeling into a number you can beat tomorrow. If writing your own questions is what stops you starting, paste a chapter into Quiz and attempt every question before checking a single answer.

Turning passive notes into recall prompts

Most students' notes cannot be used for recall because they are written as answers. Convert every heading into a question with a countable answer.

• Weak: "The League of Nations." Too big, no way to mark it. Better: "Why did the League of Nations fail? (4 reasons)"
• Weak: "Osmosis." Better: "Define osmosis using the words partially permeable and water potential."
• Weak: "Quadratics." Better: "When do I complete the square instead of using the formula? (2 situations)"

Two rules I stick to. Every prompt gets an expected number of points in brackets, so you know when you have finished rather than trailing off. And if a prompt takes longer than about 90 seconds, split it, because long prompts let you drift into rereading inside your own head. Then spread the prompts over days rather than one sitting; Flashcards handles the spacing so you are not deciding daily what to review.

Recall in the shape the mark scheme wants

Plain recall gets facts back, but marks need those facts in a particular form, so make your recall exam-shaped from the start.

• For Cambridge IGCSE and GCSE "explain" questions, recall in point-plus-development-plus-link chains, not lists of nouns. If your page says "reparations", push it to "reparations, which meant Germany's economy struggled, which fed resentment towards the settlement."
• For AQA, Edexcel and OCR sciences, recall the command word too. "State" wants one line; "explain" wants a because.
• For AP free response, retrieve a claim, then the evidence, then the reasoning that links them, in that order.
• For CBSE, recall the exact keyword terms, since answer writing is credited heavily on terminology.
• For WAEC and WASSCE theory papers, recall in numbered points, since that is how you will write them.

Once a week I do this with the clock running using Mock exam. Recall under time pressure is a different skill from recall on a quiet Sunday, and only one gets marked.

Five mistakes I made when I switched

1. Recalling immediately after reading. You are pulling from short-term memory, which flatters you. Leave ten minutes at minimum, ideally until the next day.
2. Peeking. One glance at the notes and the exercise quietly converts back into rereading. Turn the page over.
3. Only recalling definitions. They are easy to write prompts for, so they get tested to death while processes and links go unpractised. Half your prompts should be "why" or "how".
4. Not marking honestly. Unmarked recall bakes in whatever you half-remembered wrongly, and you reproduce it in the exam confidently.
5. Doing it once and calling it done. One retrieval is a start; the gains come from three or four spaced ones.

To be fair to rereading, the first pass has to be reading. If a topic is genuinely new, recall is pointless — you cannot retrieve what was never there. Read it once, make sure you understand it, and if the textbook is dense get it put another way with Explain. After that, never read it passively again.

Test yourself

1. Why do students who reread consistently overestimate how much they know?
2. In the blank page method, what should you do straight after scoring 6 out of 14 on a first attempt?
3. Rewrite this weak prompt so it can be marked: "Photosynthesis."

FAQ

Is highlighting ever worth doing?

Only as a sorting tool. Highlighting so you can later turn those lines into recall prompts is fine. If highlighting is the last thing you do to a page, you spent the time colouring.

How long should one active recall session be?

Shorter than you expect. Twenty to thirty minutes of real retrieval is harder work than an hour of reading, and quality drops fast once you tire. Two 25-minute sessions on different days beat one 50-minute session.

What if I get almost nothing on my first blank page?

That is normal, and it is information rather than failure. A first score of 3 out of 14 means the topic needs a proper read before recall, not that recall fails for you. Read it once more, wait, then try again the next day.

Does active recall work for maths as well as content subjects?

Yes, but the prompt changes. Recall methods instead of facts: cover a worked solution and reproduce it, or name which method a question type needs before doing any working. Redoing a question whose solution you just read is the maths version of rereading.

In short: rereading builds familiarity, active recall builds memory, and only one turns up in the exam hall. Read new material once, then spend the rest of your time closing the book and writing what you remember on a blank page, marked out of a fixed total. It feels worse. It works better.