Active recall: what it is, why it works, and how to do it without cheating

Rereading a chapter feels like learning. The second pass is smoother than the first, the sentences are familiar, and by the end you have the distinct sense of having got it. Then the test comes, or the conversation, and the sense turns out to have been the feeling of recognising the words rather than the ability to produce the idea.

In 2006, Roediger and Karpicke ran a study that has since been repeated many times in many forms. Students read a passage; some then reread it, others took a short test on it with no feedback. On a test five minutes later, the rereaders did slightly better. On a test a week later, the group that had been tested remembered more, and by a wide margin. The rereaders had also rated themselves as more confident. They were wrong about themselves, and the reason they were wrong is the whole subject of this article.

Active recall is the name for making yourself produce the answer from memory instead of reading it again. It is the same thing as retrieval practice, which is what the research literature calls it. This is what it is, why the effort is the point rather than an obstacle, how to do it, and the five ways people do it wrong without noticing.

What active recall is

You have some piece of knowledge. Instead of looking at it again, you close the source and try to bring it back: state the definition, name the steps, explain the argument, answer a question about it. Then you check. That attempt, made before the check, is active recall. Everything else is the setup for it.

The contrast is with passive review: rereading, re-listening, rewatching, going over your notes, looking at a highlight. In all of these the information comes to you. In recall, you go and get it, and the going is what changes something.

A flashcard is the most familiar form, and the one most people think of, but a flashcard is only active recall if you attempt the answer before you turn it over. A blank page and the instruction "write down everything you know about X" is active recall. Explaining a chapter to someone who has not read it is active recall. A practice exam taken cold is active recall. A flashcard you flip immediately is rereading in a small font.

The useful part is the attempt made before the answer is revealed.

Why it works

The effect Roediger and Karpicke measured is called the testing effect, and the plainest account of it is that memory is strengthened by use, and retrieval is the most demanding use there is. Reading a fact again asks nothing of the memory that holds it. Producing the fact from a cue asks the memory to do its job, and a memory that has done its job is more available the next time.

Two refinements matter in practice.

The first is that the difficulty is the point. Robert Bjork's phrase for this is desirable difficulties: conditions that make learning feel harder and slower in the moment, and produce more durable knowledge as a result. Rereading is easy and feels good and does little. Recall is effortful and feels bad, especially when it fails, and does a great deal. A retrieval that came easily strengthened the memory a little; one that took real effort strengthened it a lot. This is why the feeling of fluency is such a poor guide to what you know, and why the students in 2006 were confidently wrong.

The second is that it holds up against methods that look more sophisticated. Karpicke and Blunt, in 2011, compared retrieval practice with concept mapping, a technique where students draw the relationships between ideas as a diagram. Concept mapping is elaborate and active in the ordinary sense of the word; it produces a visible artefact. The students who instead spent the time recalling the material from memory did better a week later, on both factual questions and questions that required inference. When Dunlosky and colleagues reviewed ten study techniques in 2013 for the Association for Psychological Science, practice testing was one of two rated as high utility. Rereading and highlighting were rated low.

None of this means recall is the only thing that matters. Understanding something in the first place still has to happen, and a recall attempt at a thing you never understood produces a confident wrong answer. Recall is what you do with something once you have it, so that you keep it.

How to do it

Close the source and write

The simplest form and the one to start with. After reading a section, close the book and write down what it said, from memory, in your own words. Then open the book and compare. The gaps are what you did not learn; the mistakes are what you learned wrong. Both are more useful than any highlight.

This takes a few minutes per section. It is slow relative to reading. It is fast relative to reading the section again next month because you retained nothing.

Ask questions, not summaries

A summary is a description of the material. A question is a demand on your memory of it. When you turn a passage into something to review later, write the question, and write it so that a person who only recognised the topic could not answer it.

Take a passage explaining that the ribosome reads messenger RNA three bases at a time, each triplet specifying one amino acid, and that the same triplet always specifies the same amino acid across nearly all life.

A card that asks "What does the ribosome do" can be answered by anyone who remembers the word ribosome and the word protein. A card that asks "How many bases does the ribosome read per amino acid, and what does it mean that the code is nearly universal" cannot be answered without the mechanism, and the mechanism is the thing you would want to keep.

Make the front demand the back

The front of the card should be a cue that leads to exactly one answer, and the answer should require recall rather than recognition. "Describe the French Revolution" leads to a hundred answers and no way to grade yourself. "Why did the calling of the Estates-General in 1789 turn into a constitutional crisis within weeks" leads to one, and you either have it or you do not.

Multiple choice, unless it is very well written, is a recognition task: you look at four options and one of them looks right. It has its place in a well-designed exam. As a self-study tool it is a way of feeling tested without retrieving anything.

Explain it to someone

Teaching is active recall with an audience that asks follow-up questions. Explaining a chapter to a friend, or to a notebook as if to a friend, forces you to produce the argument in order, and the places where you stall are the places you did not understand. This is often called the Feynman technique, after a story about the physicist; whatever its provenance, it works because it is retrieval with a fluency check built in.

Space it

A single recall attempt strengthens a memory. A recall attempt made when the memory has begun to fade, and then another after a longer gap, strengthens it far more per minute spent. This is the spacing effect, and it is why active recall on its own is not a system. The two together are: recall gives you the strengthening, spacing tells you when to do it. Spaced repetition is the schedule; this article is what to do when the schedule says a card is due.

What is not active recall

Most of the ways people do this wrong are ways of doing something that looks like recall and skips the retrieval. In rough order of how often it happens:

Reading the answer before trying. The card comes up, you glance at the front, you flip it, you nod. This is rereading. Whether you knew the answer is unknowable, because you never asked yourself. The fix is a rule that the attempt happens out loud, or in writing, before the flip, every time, even when it feels obvious.

Grading recognition as recall. You see the answer and it looks right, so you mark the card as known. But "looks right" is recognition. The question is whether you could have produced it. If you could not, the honest grade is a miss, and the honest grade is what makes the schedule work.

Reviewing notes. Going over your notes is reading your notes. It has the same effect on retention as reading anything else you wrote, which is not much. Cover the notes and reproduce them, then compare.

Cards that are too big. "Explain the causes of the First World War" is a chapter, not a card. When a card requires five minutes and a paragraph, the recall becomes so heavy that you start skipping it or half-doing it. Split it into the five things you actually wanted and make each one a card.

Cards about the wrong thing. Dates, names, numbers, because they are easy to phrase as questions. The argument, the distinction, the reason something happened, because they are hard. After a month you know when the treaty was signed and cannot say what it changed. This is a card-writing problem, and it is the reason writing cards well is real work rather than a transcription task.

Active recall for reading, not for exams

Most of what is written about this technique is written for students with an exam in six weeks. If you are an adult reading books and papers and articles because you want to keep what is in them, the method is the same and the calibration is different.

You want fewer cards. A textbook chapter for an exam might produce thirty; a serious non-fiction book you read for its argument should produce a dozen, and the dozen should be the argument, not its furniture. You want cards about ideas, distinctions and mechanisms more than about facts, because facts can be looked up and ideas cannot. And you want the cards to carry where they came from, because in six months you will want to check whether the number on the card was the author's or yours.

You also want a daily session small enough to survive not being a student. Twenty minutes a day is a habit a person with a job can keep for a fortnight. Seven minutes is one they can keep for a year.

Tools

Any tool that shows you a question and hides the answer will do for the recall itself. Where they differ is in who writes the questions and what happens to the schedule when you miss a week. As of September 2026:

The honest general advice is that the tool matters less than the two habits: attempt before you look, and grade yourself as if someone else were checking.

Frequently asked questions

Is active recall the same as retrieval practice? Yes. Retrieval practice is the term in the research literature; active recall is the term that spread through student communities. They describe the same act: producing an answer from memory before checking it.

How is it different from spaced repetition? Recall is what you do; spacing is when you do it. Each works on its own and they work much better together. Most software that calls itself a spaced repetition system is really a scheduler for recall attempts.

Does it work for understanding, or only for facts? Both, on the evidence. Karpicke and Blunt's 2011 study tested inference questions as well as factual ones, and retrieval practice won on both. The caveat is that you cannot recall what you never understood; recall keeps understanding, it does not create it.

How often should I do it? Once at the point of learning, and then on a spaced schedule that starts at about a day and grows. Doing it every day on the same material is more than necessary; doing it once and never again is less.

Can software do the recall for me? No, and a tool that implied otherwise would be selling something else. Software can write the questions, schedule them, and hide the answers. The retrieval has to happen in your head or nothing happens at all.

What to do now

Take the last thing you read and, without opening it, write down what it said. Compare. Then turn the two or three gaps into questions whose answers you would want to have, and put them somewhere that will ask you again in a couple of days. If writing the questions is the step you know you will skip, have them proposed from the source and keep the ones you want, and spend the time on the attempts instead.