The Testing Effect
Retrieving a memory strengthens it more than restudying it does — a test is a learning event, not just a measurement. · 10 min
You have just finished a chapter. There are two ways to spend the next ten minutes: read it again, or close the book and write down everything you can remember. The second option feels worse. You stumble, you leave gaps, and there is no page to reassure you. A century of research says that this stumbling, gap-riddled ten minutes will do more for your memory than another smooth pass over the text. This folio lays out the evidence — and shows you how to collect the benefit on purpose.
Guess before you learn
Two groups study the same short passage. One group rereads it three extra times. The other group closes the page and writes out what it remembers, three times, with no rereading. A week later, everyone takes the same recall test. Who remembers more?
In Roediger and Karpicke's 2006 experiment, the rereading group recalled about 40% of the passage's ideas after a week; the repeated-recall group kept about 61%. If you picked the rereaders, you are in good company — the participants themselves predicted the same thing, and they were wrong too.
9–12
3–5
Flash cards work because of the trying, not the checking. When you look at a question and pull the answer out of your memory, that memory gets stronger — noticeably stronger than if you had just read the answer again. Scientists call this the testing effect. A test is not only a way to measure what you know. It is a way to grow what you know.
This is why flipping the card too fast wastes the trick. The strengthening happens in the seconds while you are trying to recall the answer and it has not arrived yet.
6–8
Retrieving a memory changes it. Every time you pull a fact back out — without looking — that fact becomes easier to retrieve next time and slower to fade. Psychologists call this the testing effect, or retrieval practice. In Roediger and Karpicke's 2006 study, college students read short passages, then either restudied them or repeatedly recalled them. Five minutes after the last session, the restudiers were ahead. One week later, the recallers were far ahead: about 61% of the passage's ideas kept, against about 40%.
Notice both halves of that result. Rereading genuinely is better for a test in five minutes — which is why it feels effective. Retrieval wins on every horizon that matters.
9–12
The effect is large by the standards of education research. A 2017 meta-analysis by Adesope and colleagues, pooling hundreds of comparisons, put the advantage of practice testing near g = 0.61 — meaning the average tested student lands about six-tenths of a standard deviation above the average restudier. Few study techniques come close to that figure.
The uncomfortable half of the finding is metacognitive: Roediger and Karpicke's restudiers predicted they would remember more, and the recallers predicted less — the exact reverse of the outcome a week on. Fluent rereading inflates your judgment of learning (folio 4); effortful retrieval deflates it. How a study method feels is not evidence about how it works.
K–2
You meet someone new named Ruby. Later you try to say her name from memory — and you get it. Trying to remember the name is what makes it stay.
Hearing the name again helps a little. Pulling it out of your own head helps much more. Try first, then check.
Undergrad
Why does retrieval strengthen memory more than re-exposure? Current accounts converge on effort at the right site. The elaborative-retrieval view holds that searching memory activates related knowledge and lays additional routes to the target. The episodic-context account (Karpicke, Lehman, and Aue, 2014) holds that retrieval reinstates and updates the memory's temporal context, making it reachable from more future states. On either account, the difficulty is the mechanism, not a by-product.
Bjork's distinction between retrieval strength (current accessibility) and storage strength (durable learning) organizes the data: restudy raises retrieval strength briefly; successful retrieval raises storage strength — and the harder the successful retrieval, the larger the gain.
Postgrad
Meta-analytically the effect is robust: Rowland (2014) reports g ≈ 0.50 against restudy controls, with recall formats outperforming recognition formats; Adesope, Trevisan, and Sundararajan (2017) report g ≈ 0.61 across 272 effect sizes, with benefits surviving — though shrinking — when no feedback is given. Test-potentiated learning and forward-testing effects extend the phenomenon beyond the practiced items themselves.
Boundary conditions matter for application: gains require retrieval success rates high enough to consolidate (feedback rescues low-success practice), and the advantage grows with retention interval — short-delay assessments systematically understate it, which is one reason classroom habit drifted toward rereading in the first place.
testing effect
The finding that retrieving information from memory strengthens that memory more than restudying it does. Also called retrieval practice.
Why is this true?
Why does a delayed test reveal the testing effect when an immediate test can hide it?
Restudying gives a short-lived boost that is still present five minutes later but mostly gone within days. Retrieval builds slower-fading memory, so its advantage only becomes visible once the quick boost has decayed.
The recall group started lower — about 71% at five minutes, because three of their four sessions had no page in front of them. Then the curves cross. By day seven the recallers hold 61%, twenty-one points clear of the rereaders. Same passage, same total time, different act: one group practiced putting information in; the other practiced getting it back out — the act the final test actually demands (folio 1).
To use the effect, convert exposure into retrieval. After a page, close it and say what it claimed. Turn headings into questions and answer them before reading on. Recite a definition before checking it. The Retrieval Gate — answer before you continue — that ends every section of this Archive is the same move built into the furniture: nothing about it is decoration.
One more thing the recall group collected along the way: errors. They guessed wrong, noticed, and read the correction. Folio 6 takes up those errors — why a confident miss, corrected, can become the strongest memory you make all week.
Practice — new ink and old, interleaved
1.Right after a fourth smooth rereading, your judgment of learning is high. What is that feeling actually tracking?
2.Rereading a chapter mostly exercises which act of memory?
3.A list took 40 minutes to learn and 30 minutes to relearn a week later. What is the savings, in percent?
4.Turn this highlighted sentence into a self-test question: 'The hippocampus replays the day's learning to the cortex during slow-wave sleep.'
5.Match each term to its description.
6.You just read that the mitochondrion produces the cell's usable energy. Write a card question whose answer is 'usable energy'.
7.From folio 3: forgetting is steepest —
8.A meta-analysis reports retrieval practice at g ≈ 0.61. What does that number mean?
9.From memory: give one reason retrieval strengthens memory more than rereading does.
Retrieval is effortful reconstruction — searching memory activates related knowledge and updates the memory's context, so the fact becomes reachable from more future situations.
How close were you? Grade yourself honestly — it sets your review date.
10.By the plate above, about what percentage savings remained one full day after learning?