The Revision That Feels Like Failing
Five minutes after they finished, the students who had just re-read their notes were ahead. A week later they had lost badly. The odd part is that nearly every teenager now knows which method works, and the reason they still avoid it isn't laziness.
Five minutes after they had finished, the students who had simply read the passage over and over were winning. They could recall 83 per cent of it. The ones who had shut their notes and tried to write down what they remembered managed 71 per cent.
Then both groups were asked how much they thought they would still know a week later. The re-readers were the confident ones. A week on, they remembered 40 per cent. The group who had tested themselves remembered 61 per cent.
Henry Roediger and Jeffrey Karpicke ran that experiment at Washington University in 2006, and versions of it have been repeated many times since. It is one of the sturdier findings in the study of how people learn. It also explains why revision in most houses goes the way it does.
Because the striking part is not that self-testing won. It is the five-minute result. For a little while, re-reading really was better. It felt better too, and the feeling outlasted the advantage by a long way.
When you read a page for the fourth time it goes down easily. The words are familiar. Nothing snags. The brain reads that smoothness as knowing, and it is a fair mistake to make, because most things we genuinely know do feel easy.
But familiarity and memory are not the same thing. Recognising a sentence while it sits in front of you is a different job from producing it in an exam hall when it doesn't.
Hauling something out of your own head without the notes is the harder job, and it is the one that builds the memory. Each time a child finds an answer themselves, the route back to it gets easier to find. Re-reading never asks them to look for the route at all.
So the revision that looks most productive often isn't. William Wadsworth surveyed nearly 40,000 secondary pupils across 56 English schools about how they actually revise. Re-reading, highlighting and making notes took up 37 per cent of their revision time. Even flashcards get misused: around 40 per cent of the pupils using them spent most of their time making the cards or reading through them rather than covering the answer. And most were cramming, with three-quarters not spreading the work across days.
Knowing did not translate. Some 56 per cent said they tested themselves often, and 41 per cent spaced their revision.
What they said about why is the part worth sitting with. The researchers took the students who rated these methods as highly effective, nine or ten out of ten, and asked why they still avoided them. Up to half said it took too long. Up to 41 per cent said it was too much effort. And as many as 47 per cent said it made them anxious.
That last one is not irrational. Testing yourself means finding out, over and over, how much you don't know yet. A child with three weeks until an exam and a stomach full of dread does not want an hour of being shown the gaps. Re-reading offers the opposite: proof every few minutes that they know things. It is the worse method and the kinder one.
In practice the better version is less elaborate than it sounds. Shut the book, write down everything you can remember about photosynthesis, then open it and see what is missing. Answer the questions at the end of the chapter without flicking back. Get someone to read out the French vocabulary in English. The difference between a good revision hour and a wasted one is often just whether the book was open.
The Education Endowment Foundation went through 295 studies on cognitive science in classrooms. It found the idea broadly supported, but noted that much of the work was done by researchers running tightly scripted lessons, and that the clearest gains are in remembering facts and vocabulary. Whether it helps as much with the kind of question that asks a child to weigh two things against each other is less settled.
A trial published last year makes the point neatly. Researchers taught atomic structure to 336 pupils across six non-selective state schools in England, using a single spaced-out hour in place of the ten or twelve hours the topic normally takes. On its own, that lesson gave a bump at one week which had gone by two months. Where it was added on top of ordinary teaching, pupils were doing 50 to 90 per cent better than the others at the later test.
So it does not stand in for being taught properly. It is the thing that makes the teaching stay put.
What it does change is a conversation most of us have had. A child comes downstairs after an hour with their books and we ask how it went. If they say it was fine and they knew most of it, that may mean very little. If they say it was horrible and they couldn't remember anything, something probably happened.
The discomfort is not a sign that revision is going badly. Often it is the only evidence that any is going on. Which makes "how did it go?" close to useless as a question, and leaves us with the harder job: persuading a tired fifteen-year-old to keep choosing the thing that feels worse.
Then both groups were asked how much they thought they would still know a week later. The re-readers were the confident ones. A week on, they remembered 40 per cent. The group who had tested themselves remembered 61 per cent.
Henry Roediger and Jeffrey Karpicke ran that experiment at Washington University in 2006, and versions of it have been repeated many times since. It is one of the sturdier findings in the study of how people learn. It also explains why revision in most houses goes the way it does.
Because the striking part is not that self-testing won. It is the five-minute result. For a little while, re-reading really was better. It felt better too, and the feeling outlasted the advantage by a long way.
The Five Minutes That Fool Everyone
Most of us can picture the scene. A child at a desk with a highlighter, working steadily through a textbook, pages turning. Four colours on every page by the end. It looks exactly like studying, which is the problem.When you read a page for the fourth time it goes down easily. The words are familiar. Nothing snags. The brain reads that smoothness as knowing, and it is a fair mistake to make, because most things we genuinely know do feel easy.
But familiarity and memory are not the same thing. Recognising a sentence while it sits in front of you is a different job from producing it in an exam hall when it doesn't.
Hauling something out of your own head without the notes is the harder job, and it is the one that builds the memory. Each time a child finds an answer themselves, the route back to it gets easier to find. Re-reading never asks them to look for the route at all.
So the revision that looks most productive often isn't. William Wadsworth surveyed nearly 40,000 secondary pupils across 56 English schools about how they actually revise. Re-reading, highlighting and making notes took up 37 per cent of their revision time. Even flashcards get misused: around 40 per cent of the pupils using them spent most of their time making the cards or reading through them rather than covering the answer. And most were cramming, with three-quarters not spreading the work across days.
They Already Know. They Just Don't Want To.
The obvious fix would be to tell children about this. It has been tried. Shana Carpenter and Christopher Sanchez asked 185 university students in Oregon what they knew about study strategies. Almost all had heard of testing yourself: just 1 per cent said they had never come across it. Only 3 per cent had never heard of spreading revision out over time.Knowing did not translate. Some 56 per cent said they tested themselves often, and 41 per cent spaced their revision.
What they said about why is the part worth sitting with. The researchers took the students who rated these methods as highly effective, nine or ten out of ten, and asked why they still avoided them. Up to half said it took too long. Up to 41 per cent said it was too much effort. And as many as 47 per cent said it made them anxious.
That last one is not irrational. Testing yourself means finding out, over and over, how much you don't know yet. A child with three weeks until an exam and a stomach full of dread does not want an hour of being shown the gaps. Re-reading offers the opposite: proof every few minutes that they know things. It is the worse method and the kinder one.
In practice the better version is less elaborate than it sounds. Shut the book, write down everything you can remember about photosynthesis, then open it and see what is missing. Answer the questions at the end of the chapter without flicking back. Get someone to read out the French vocabulary in English. The difference between a good revision hour and a wasted one is often just whether the book was open.
It Is Not a Magic Trick
This is where the research deserves some honesty, because retrieval practice has become something of a slogan in English schools and the evidence is narrower than the enthusiasm.The Education Endowment Foundation went through 295 studies on cognitive science in classrooms. It found the idea broadly supported, but noted that much of the work was done by researchers running tightly scripted lessons, and that the clearest gains are in remembering facts and vocabulary. Whether it helps as much with the kind of question that asks a child to weigh two things against each other is less settled.
A trial published last year makes the point neatly. Researchers taught atomic structure to 336 pupils across six non-selective state schools in England, using a single spaced-out hour in place of the ten or twelve hours the topic normally takes. On its own, that lesson gave a bump at one week which had gone by two months. Where it was added on top of ordinary teaching, pupils were doing 50 to 90 per cent better than the others at the later test.
So it does not stand in for being taught properly. It is the thing that makes the teaching stay put.
What it does change is a conversation most of us have had. A child comes downstairs after an hour with their books and we ask how it went. If they say it was fine and they knew most of it, that may mean very little. If they say it was horrible and they couldn't remember anything, something probably happened.
The discomfort is not a sign that revision is going badly. Often it is the only evidence that any is going on. Which makes "how did it go?" close to useless as a question, and leaves us with the harder job: persuading a tired fifteen-year-old to keep choosing the thing that feels worse.
Sources & Further Reading