Skip to the content

Retrieval practice: what it is, how to use it, and what the research says

Having students pull what they've learned back out of memory, by quizzing, recalling or explaining, instead of reading it again.

Evidence StrongArea How students learn and remember6 sources, about 10 minutes to readChecked October 2026

What the research says

  1. Strong
  2. Promising
  3. Mixed
  4. Thin
  5. Contradicted

Retrieval practice is one of the best-supported techniques in the science of learning. Students remember more when they pull information out of memory than when they read it again. A review of 50 experiments run in real classrooms found the benefit across education levels, subjects and test formats, and a federal practice guide rates quizzing to bring key content back as Strong. The gain shows up days and weeks later, not five minutes later, which is why it can feel less productive than rereading.

Rated Strong. Four syntheses and one lab or classroom experiment support it. One review or weaker study adds context but doesn't move the rating. How we rate the evidence

Checked against each source in October 2026.

The studies behind this rating

  1. Agarwal et al. (2021)Systematic reviewSupports

    Counting only experiments run in real classes, retrieval practice improved learning across education levels, subjects, test formats and delays, and 57% of the effect sizes were medium or large. Only 6% of the experiments were run outside Western, educated, industrialized, rich and democratic countries.

    Sample 50 classroom experiments, 49 effect sizes, 5,374 students, from nearly 2,000 abstracts screenedWho School and collegeRead it at the source (opens in a new tab)
  2. Adesope et al. (2017)Meta-analysisSupports

    Across the studies pooled, taking practice tests helped learning more than restudying and more than every other comparison condition. How much it helped depended on the kind of practice test, the learners, the outcome measured and how the study was run.

  3. Pashler et al. (2007)Practice guideSupports

    Recommendation 5b: Use quizzing to promote learning: quizzes to re-expose students to key content. Panel's rating: Strong.

    A What Works Clearinghouse panel made seven recommendations for organizing teaching and study. It rated the evidence Strong for using quizzes to bring key content back, and for asking deep explanatory questions; Moderate for spacing learning over time, interleaving worked examples with problems, pairing graphics with words, and linking abstract and concrete ideas.

  4. Rowland (2014)Meta-analysisSupports

    Pooling studies that compared taking a test with restudying, the review found support for effortful processing as one reason testing helps: recall tests produced larger benefits than recognition tests such as multiple choice. Explanations based on richer meaning alone were only weakly supported.

  5. Roediger & Karpicke (2006)Lab experimentSupports

    Students who took recall tests on a passage, without feedback, remembered much more of it two days and a week later than students who spent the same time rereading it. Rereading won on a test five minutes later, and it made students more confident, even though it helped them less over time.

    Sample Two experiments with students reading prose passagesWho College studentsRead it at the source (opens in a new tab)
  6. Dunlosky et al. (2013)Review (adds context, doesn't move the rating)Supports

    Rated practice testing high utility, one of only two techniques out of ten to get the top rating, because it helped learners of different ages on many kinds of test, including in real classes. Rereading and highlighting were rated low.

    Sample Ten study techniques, each reviewed across learners of different ages, materials and testsWho School and collegeRead it at the source (opens in a new tab)

What it is

Retrieval practice means asking students to recall what they've learned, from memory, before they look it up. A quick quiz, a brain dump on a blank page, flashcards answered before the flip, a partner asking questions, or a "what do you remember from yesterday?" warm-up are all retrieval. The point isn't the score. It's the act of bringing the memory back.

Researchers call it the testing effect, because most of the early studies used tests. The word worries teachers who already feel buried in testing, so it helps to keep the two apart. Retrieval practice is short, frequent and low-stakes or no-stakes, and it's there for learning, not for the gradebook. A five-question warm-up that nobody records is retrieval practice. A unit test that only measures is something else, even though it helps memory too.

It pairs naturally with two other habits: spacing (coming back to material over days and weeks instead of in one block) and feedback (showing the right answer after students have tried). Together they make the review part of a lesson do far more work than a reread of the notes.

What counts, and what doesn't

  • Counts: free recall (write everything you remember), short-answer questions, flashcards answered before turning them over, explaining a process to a partner without notes, labeling a blank diagram, solving a problem type from memory.
  • Doesn't count: rereading notes, highlighting, copying from the board, watching the video again, or looking at the answer while "practicing". These feel productive. The review by Dunlosky and colleagues rated rereading and highlighting low utility.

Why it works

Every time a student retrieves something, it gets easier to bring back the next time. Rereading puts the information in front of the student again, but it doesn't make them practice what they'll actually need to do later: bring it to mind without the page in front of them.

The best-known demonstration is a 2006 experiment by Henry Roediger and Jeffrey Karpicke. College students read short prose passages, then either reread them or took recall tests with no feedback. On a test five minutes later, rereading won. Two days and a week later, the students who had been tested remembered substantially more. The rereaders were also more confident they'd remember, and that's the trap: rereading feels like learning because the text feels familiar.

Effort seems to be part of it. A meta-analysis by Christopher Rowland found support for effortful processing as one reason testing helps, and found bigger benefits from recall tests (write the answer) than from recognition tests (pick the answer). That's a practical point for the classroom: questions that make students produce an answer are worth more than ones that let them spot it.

Retrieval also shows students and teachers what's missing. A student who can't recall a step finds out now, not on the unit test, and a teacher scanning the answers sees which idea to reteach tomorrow. That's where retrieval practice overlaps with formative assessment.

Why it can feel worse at first Retrieval is harder than rereading, and students get things wrong while they do it. Expect some groaning, keep the stakes at zero, and give the right answers straight after.

How to start, by grade

Grades K–2

  1. Start the day with "What do you remember from yesterday?" Give thirty seconds of quiet thinking, let partners tell each other, then take a few answers.
  2. Use pictures, sounds and objects instead of writing: hold up a letter and ask for its sound, show a shape and ask its name, say a word and ask for a rhyme.
  3. Play quick whole-class recall: thumbs up if it's true, show the answer on your fingers, whisper it to your neighbor. Then pick a name with the random name picker so everyone expects to be asked.
  4. Bring back last week's and last month's learning, not only yesterday's. Mix two or three old items into every warm-up.
  5. Always say or show the right answer straight after, so nobody walks away with a wrong one.

Grades 3–5

  1. Open lessons with a three-to-five-question warm-up on earlier material: one question from yesterday, one from last week, one from last month.
  2. Teach the brain dump: two minutes to write everything they remember about a topic on a blank page, then compare with a partner, then open notes and add what they missed in a different color.
  3. Use flashcards the right way: say or write the answer before turning the card, and keep the missed cards in play until they're right twice.
  4. End some lessons with an exit ticket that asks students to recall something, not just rate how well they understood.
  5. Before a test, run a review game with the review game board maker, and have every student write an answer on paper or a whiteboard, not only the team captain.

Grades 6–8

  1. Put a weekly no-stakes quiz in the timetable. Mix questions from across the unit and from earlier units, and let students mark their own.
  2. Teach students how memory works: show them the difference between rereading and self-testing, and ask them to plan their own study around practice questions.
  3. Use free recall around the lesson: two minutes before it to write what they already know, two minutes after it to write what they learned, without looking.
  4. Ask for explanations as well as facts. "Why does this work?" and "How would you explain this to a younger student?" are retrieval too.
  5. Use the classroom quiz show maker for team review, with every student writing an answer before the team commits to one.

Grades 9–12

  1. Start each class with two or three recall questions from earlier units, so old material keeps coming back all year.
  2. Replace a review lecture with a practice test students answer first and then check against the answer key.
  3. Give students a study plan built on self-testing for the week before an exam, and ask them to show you their practice questions.

Examples

  • The three-question warm-upGrades 3–8

    On the board as students come in: a question from yesterday, a question from last week, a question from earlier in the year. Students answer on a slip or in a notebook, without notes, for four minutes. You read the answers aloud, they check their own, and you note which question most people missed. Nothing is graded; the missed question goes into tomorrow's lesson.

  • Brain dump with a second colorGrades 2–8

    "Write down everything you remember about the water cycle. Words, pictures, arrows, anything. Two minutes, no notes." Partners compare and add. Then everyone opens their notes and adds the missing pieces in a different color. The second color shows each student exactly what they'd forgotten, and shows you what the class forgot.

  • Letters and sounds, three minutesKindergarten–1st grade

    Flash a letter card and the class says its sound together. Then pick one name with the picker and flash another. Then mix in two letters from last week and one from last month. Every child retrieves a dozen times in three minutes, and you hear who hesitates.

  • Flashcards that actually testGrades 2–12

    Students make cards with a question on one side and the answer on the other. They answer out loud or on scrap paper before they flip. Cards they get right go in one pile; cards they miss go back in the deck. They stop when every card has been right twice, not when they've looked at each card once.

  • Review game where everyone answersGrades 3–12

    A board-style review game on the projector, but before any team answers, every student writes their own answer on a whiteboard and holds it up. You see the whole room's answers at once, and nobody coasts while the team's strongest student answers for them.

Common mistakes

Grading every quiz.
Retrieval practice works as practice. When every quiz counts, students cram, copy and worry, and you drown in marking. Keep most of it ungraded and use what you see to plan what to reteach.
Only ever asking about yesterday.
Memory fades over weeks, not hours. Mix in older material on purpose. Spreading practice out over time is itself well supported: the Dunlosky review rated distributed practice high utility, and the IES guide rates spacing learning over time Moderate.
Letting the fastest hands answer.
If three hands go up and you call on one student, the rest of the room didn't retrieve anything. Have everyone write, show or tell a partner first, then call on names at random.
Using only multiple choice.
Picking an answer is easier than producing one, and Rowland's meta-analysis found larger benefits from recall formats than from recognition formats. Use short answers, fill-in-the-blank and "write what you remember" as well.
Skipping the answers.
Always go over the answers straight away, or let students check against their notes. Students need to know which answers were wrong, or they have no way to fix them.
Calling it a test.
Call it a warm-up, a brain dump or a check. The name changes how students feel about it, and the stakes should match the name.

Time and cost

Time: three to ten minutes, a few times a week. It replaces some of the review you already do rather than adding to it.

Money: nothing. Paper slips, mini whiteboards or the free tools below are enough.

Preparation: writing good questions is the real work. Keep a running list per unit and reuse it all year, since retrieval questions are meant to come back. A question bank shared across a grade team saves everyone time.

Training: very little. The main change is a habit: before you re-explain something, ask students to recall it first.

Debates and open questions

Does it hold up in real classrooms, not just the lab? The famous experiments used college students and short passages. Agarwal and colleagues set out to answer this by counting only experiments run in real classes, and found the benefits held up across education levels and subjects. Their review notes that only a small share of those experiments came from outside Western, educated, industrialized, rich and democratic countries, so we know less about other settings.

Is it only good for memorizing facts? Many of the best-known studies measure recall of facts and concepts. Retrieval questions can also ask students to explain, apply and compare, and the same IES guide separately rates "ask deep explanatory questions" as Strong. How far quizzing on facts carries over to deeper understanding depends on what you ask.

How big is the effect? We don't quote a single headline number. Meta-analyses report average benefits over restudying, but Adesope and colleagues found the size depends on the kind of practice test, the students, the outcome measured and how the study was run. Treat any one figure with care, including ones you see on posters.

What about anxious students? Some teachers worry that frequent quizzing raises anxiety. Keeping it ungraded, short and private (students check their own answers) is the usual answer. We haven't yet added a study on anxiety to this card; when we do, it will be listed above.

Tools that do it

Free, on this site, nothing to sign up for. Class lists stay in your browser.

  • Flashcard maker: two-sided cards from your own list; students answer before they flip, and the missed cards stay in the pile.
  • Review game board maker: a board-style review game for the projector. Mix questions from this week and from earlier units.
  • Classroom quiz show maker: teams answer on the big screen; have every student write their own answer before the team commits.
  • Exit tickets: printable slips. Write a recall question on them, not "How well did you understand today?"
  • 3-2-1 exit slip: three things learned, two questions, one thing to remember: a ready-made brain dump.
  • Random name picker: call on names at random after everyone has thought or written, so every student retrieves, not just the volunteers.
  • Exit ticket wheel: spin for an end-of-lesson recall prompt.
  • 30 exit ticket ideas: prompts that work as retrieval questions.

Questions teachers ask

Is retrieval practice the same as testing?
Not in the way teachers usually mean. It uses questions, like a test, but it's frequent, short and low-stakes or ungraded, and its job is to strengthen memory rather than to measure it. Graded tests help memory too, but they bring stress and marking that retrieval practice doesn't need.
How often should we do it?
Most teachers who use it build in a few minutes several times a week, usually at the start of lessons. The research doesn't give one right number. What matters more is coming back to the same material over days and weeks, rather than once.
Does it work with young children?
The classroom review by Agarwal and colleagues found benefits across education levels. With young children, retrieval is oral, quick and playful: letter sounds, names of shapes, number facts and "what happened in the story?" all count, as long as children answer from memory before you tell them.
Is multiple choice okay?
Yes, as one kind of question. Recall formats (short answer, fill in the blank, write what you remember) produced larger benefits than recognition formats in Rowland's meta-analysis, so don't rely on multiple choice alone. When you do use it, make the wrong options plausible so students have to think.
Should I grade it?
Mostly no. A grade turns practice into a performance, and students start protecting the grade instead of finding their gaps. Use the answers to decide what to reteach, and keep grades for the assessments that are meant to measure.
How is this different from formative assessment?
They overlap. Retrieval practice is about strengthening memory; formative assessment is about using what students show you to change what you teach next. A good warm-up does both: students retrieve, and you see what to reteach. The formative assessment card covers the second half.

Cite this card

Peabody Teach. (2026, October 29). Retrieval practice: what it is, how to use it, and what the research says. https://teach.peabody.kids/library/practices/retrieval-practice

Written for teachers by Peabody Teach from the sources listed above, which we read on the publisher's or clearinghouse's own page. We summarize and link; we don't copy. No ads, no affiliate links. Something wrong or out of date? Tell us.