steady

Memory and learning · Learning it

Test yourself, don't reread

5 minEvidence: solid

Do this

Close the notes. Write down what you remember. Then check. Once now, once tomorrow.

How to do it

  1. Close the notes first. Recall with nothing in front of you; that's the part that works.
  2. Write it down or say it out loud. Then you know what you actually got, not what you think you got.
  3. Then check. The studies worked without feedback; checking just catches what you got wrong.
  4. Again tomorrow. Spreading sessions out was the other technique the big review rated highest.

Why recall beats rereadingTap a word for what it means

  1. Retrievalpulls it back out

    Recalling means searching for it with nothing in front of you. In one pooled analysis producing the answer beat recognising it; its author put that down to the effort.

  2. Hippocampuswas more active

    In a scan study of 22 people, recalls that led to memory a day later went with more activity here than rereading did. Scans track blood flow, so this shows the area at work, not the cause.

  3. Forgettingslows down

    Students who'd recalled a passage once remembered as much after a week as rereaders did after two days.

  4. Youkeep more, feel less sure

    Rereaders felt more confident about the week-later test and did worse on it. The ease of rereading is what fools you.

Why it works

Rereading feels like learning. It isn't, or not much. Evidence: solid Pulling something out of your head is what strengthens it, and it's been replicated many times. Evidence: solid

Works for a report, a name, a language, a talk. Evidence: worth trying The uncomfortable feeling of not quite remembering is the exercise working. Evidence: worth trying

The long readEvery step of the chain, each study with its numbers, and what's still unproven. 3 min read.

What happens in your brain

Recall is a search. With the notes closed you have to find the thing yourself. A pooled analysis of testing studies found that recall practice, where you produce the answer, helped later memory more than recognition practice like multiple choice. Evidence: decent Its author read that as the effort of the search being part of the benefit. Evidence: decent Not every analysis agrees on which kind of practice test works best. Evidence: decent

The memory system is working harder. In a scan study of 22 people learning word pairs, the recalls that led to success on a test a day later went with more activity in the hippocampus, the brain's memory centre, and in parts of the prefrontal cortex than restudying did. The hippocampus was also more in step with other areas. Evidence: decent Scans track blood flow, so this shows those areas at work during recall, not that they cause the gain. Evidence: decent

Forgetting slows. In the best-known study, students who took one recall test remembered as much after a week as rereaders did after two days. Evidence: solid

You feel less sure, and you're wrong about it. Students who reread four times were more confident they'd remember the passage in a week than students who'd recalled it. They remembered less. Evidence: solid In a later study, students' predictions of how much they'd remember didn't track what they actually remembered at all. Evidence: solid

The proof

  • In the best-known study, 180 students read a prose passage either four times, or once and then spent the other three periods writing down what they could recall. Five minutes later the rereaders knew it better, 83% to 71%. A week later it had flipped: the recall group kept 61%, the rereaders 40%. Evidence: solid
  • In a first experiment with 120 students, one recall test instead of one more reading meant 56% against 42% a week later. Evidence: solid
  • Students learning 40 Swahili-English word pairs recalled about 80% a week later when the pairs stayed in the testing, and about a third when a pair was dropped from testing once they'd got it right, however many more times they studied it. Evidence: solid
  • A meta-analysis of practice-testing studies found practice tests beat restudying and every other comparison. Evidence: solid
  • A large review of ten learning techniques rated practice testing one of the two most useful, alongside spacing sessions out. Rereading rated low. Evidence: solid

How to do it

  • Close the notes first. Recall with nothing in front of you. Evidence: solid
  • Write it down or say it out loud, so you know what you actually got. Evidence: worth trying
  • Then check. The recall studies gave no feedback and still worked; checking just catches what you got wrong. Evidence: solid
  • Again tomorrow. Spreading sessions out was the other technique the big review rated highest. Evidence: solid

Watch out

If the test is in five minutes, rereading does just as well or better; the benefit of recall shows over days. Evidence: solid Most of the evidence is students learning word lists and short texts. It's been replicated many times, but using it for a work report or a talk is an extrapolation. Evidence: decent That the uncomfortable feeling means it's working is the usual reading, not something the studies measured. Evidence: worth trying

Arrows to move, Enter to open, Esc to close.