A retrieval practice grid is a set of short recall questions arranged by when the material was first learned (last lesson, last week, last month), often with more points for older material. It combines retrieval practice and spaced review, both strongly supported by research. The grid format itself hasn't been studied, so its details are design choices.
Key takeaways
- Retrieval practice means recalling information from memory rather than re-reading it. A meta-analysis of 222 classroom studies found quizzing raised achievement by an average of g = 0.499.
- In the classic study, re-reading beat a recall test when the final test came five minutes later, but the recall test won by 14 percentage points after two days and after a week.
- Spacing retrieval out beat massing it together (g = 0.74 across 29 studies), but expanding the gaps between reviews was no better than keeping them even.
- Classroom quizzes helped more with corrective feedback, when given after teaching, and when their format matched the final assessment.
- No published study has tested the retrieval grid itself, so treat its layout, points and timing as design decisions, and keep the format simple enough that learners think about the content, not the scoring.
Retrieval grids spread through school staffrooms because they make two good ideas easy to use at once. A single sheet asks learners to recall material from last lesson, last week and last month, and often gives more points for the older questions. Both ideas underneath it, retrieval practice and spaced review, are among the best-supported findings in learning research. The grid itself, as far as we can find, has never been tested.
That gap matters more than it sounds. The pages that rank for “retrieval practice grid” are mostly templates and classroom how-tos: colour-coded tables for a five-minute lesson starter, marked by cold-calling and a show of hands. They rarely separate what the research supports from what is a teacher’s design choice, and none of them deals with a self-paced online course, where there’s no teacher to read out the answers and no class to keep the routine going. This entry covers what the evidence says, what the grid adds, and how to build one into a course that learners work through alone.
What the research says
What retrieval practice is
Retrieval practice means recalling information from memory rather than re-reading or re-watching it. A short quiz, a blank page on which learners write what they remember, a flashcard answered before it’s turned over: all of these ask the learner to pull the information back out. The finding behind it is often called the testing effect, because the first studies used tests, but the tests don’t need to be graded or even marked by anyone else.
The classic study, and why the benefit is easy to miss
The study most often cited is by Henry Roediger and Jeffrey Karpicke (Roediger & Karpicke, 2006). In their first experiment, 120 university students read short prose passages. They then either read each passage again or took one free-recall test on it, writing down what they could remember, with no feedback. A final test came five minutes, two days or one week later.
Re-reading won at 5 minutes; a recall test won by 14 points at 2 days and 1 week
Proportion of idea units recalled from a prose passage that students had either restudied or recalled once in a test without feedback, by delay before the final test (Experiment 1, 120 undergraduates) · % of idea units recalled
At five minutes, re-reading won: 81% of the passage’s ideas recalled against 75%. After two days the tested passages were ahead, 68% against 54%, and after a week the gap held at 56% against 42%. The authors’ conclusion is the one every summary of retrieval practice repeats: “Testing is a powerful means of improving learning, not just assessing it.”
The detail that matters for course design is the first result. If you judge a lesson by how well learners do straight after it, re-reading looks better. In the second experiment, students who studied a passage repeatedly were more confident they’d remember it in a week than students who were tested, and they actually did worst. Learners, and the people who design for them, can’t reliably feel which method is working.
Does it hold up in real classrooms?
Laboratory studies with university students and short passages are a long way from a twelve-week course. The strongest answer to that worry is a meta-analysis by Chunliang Yang and colleagues that pooled only classroom studies: 222 independent studies with data from 48,478 students, from elementary school to university (Yang et al., 2021).
A g of about 0.5 is a medium effect, and it’s an average across real courses with real assessments, not a best case. The same review found that the benefit depended on how the quizzes were run. Four of its moderators translate directly into course design:
- Feedback. Quizzes followed by corrective feedback had a larger effect (g = 0.537) than quizzes without it (g = 0.374).
- Timing. Quizzes given after teaching were far more effective (g = 0.536) than quizzes given before it (g = 0.186).
- Format. Quizzes helped more when their format matched the final assessment (g = 0.531) than when it didn’t (g = 0.399).
- Repetition. The more occasions on which content was quizzed, the larger the gains.
Spacing it out
The second ingredient in a grid is spacing: returning to material after a gap rather than practising it all in one sitting. Antoine Latimier, Hugo Peyre and Franck Ramus pooled 29 studies that compared spaced and massed retrieval practice (Latimier et al., 2021). Spacing the recall out had a large benefit, with an average effect of g = 0.74 across 39 effect sizes.
They also tested a popular rule: that the gaps between reviews should get longer each time (an expanding schedule, the logic behind many flashcard apps). Across 54 effect sizes they found no significant difference between expanding and even gaps.
For a course designer, that’s good news. You don’t need an algorithm. Returning to earlier material on a regular rhythm, such as at the start of each module, captures most of what the research supports.
What the grid adds, and what’s untested
Pooja Agarwal’s write-up for RetrievalPractice.org credits the retrieval grid to Kate Jones, a teacher and writer, and describes it as “simply a table with prompts for retrieval practice”, with each box colour-coded by when the topic was first taught: the last lesson, last week, two weeks ago, and so on (Agarwal, 2018). Students answer in any order, some grids give points for how far back they can recall, and afterwards students can check their own work or get feedback from the teacher.
What the grid adds is packaging. It makes spaced retrieval a routine, and it makes the age of each question visible, so older material doesn’t quietly drop out of review. Those are useful things. But no published study we could find has compared a grid with the same questions in a plain list, or tested whether the points help. Andrew Watson, writing for Learning and the Brain, put it plainly: “I don’t know of any research on this subject” (Watson, 2020).
Watson raises one specific concern: asking learners to recall facts about a topic and keep a running points total at the same time loads working memory with two unrelated tasks. His suggested fixes are simple: check whether learners actually struggle, set a number of questions per row instead of a points target, or let them note the points as they go. That’s a reasoned argument from cognitive load research, not a tested finding, but it’s a useful warning against decorating a grid until the scoring is the hard part.
How strong is the evidence?
Strong, for retrieval practice and for spacing. Both rest on large meta-analyses, and the classroom review shows the effect holds in real courses at every level of education. The effects vary with feedback, timing and format, so the design matters. The grid format is a different matter: it’s a sensible way to package the two strategies, but it has no evidence of its own, so its layout, colours and points are choices to make for your learners, not rules to copy.
How to apply it in a course
A paper grid works in a classroom because a teacher puts it on the board, keeps time and reads out the answers. A self-paced course has none of that, so the grid needs translating. The version we’d build is a short review set at the start of each module.
We’ll use one example throughout: an online course for new team leaders, with modules on running one-to-ones, giving feedback, delegating and handling conflict. It’s an invented example, but it has what most professional courses have: ideas learners must still recall, and use, weeks after the module that taught them.
At Avondale University, we build recall questions into most of our short courses. In a governance course for board members, for example, we teach the same learning outcomes in three different ways. Learners read about a board scenario. They work through flashcards that each set out a scenario and ask them to decide whether the board or management acted correctly. And they take a quiz on a recorded lecture. Coming at the same ideas three ways, two of them by recalling and applying rather than re-reading, helps the knowledge stick.
That design matches the research above. The flashcards and the quiz are both retrieval, and in the classroom review by Yang and colleagues, the more occasions on which content was quizzed, the larger the gains (Yang et al., 2021). The scenario flashcards also ask learners to apply what they know, which is closer to how a board member will use it.
1. Open each module with a review set
Place the review at the start of the module, before the new material. It catches what learners have forgotten since last time, and it sits after the teaching it reviews, which is where Yang and colleagues found quizzes worked best. Five to eight questions is enough; it should take a few minutes, not become a module of its own.
2. Build the rows from earlier modules
This is the grid’s real contribution: questions from several points in the past, not just the last lesson. For the start of module 4 in the team leader course, the set might look like this.
| Last module (delegating) | Two modules back (feedback) | Earlier (one-to-ones) |
|---|---|---|
| What are the three things to agree before you hand over a task? | What’s the difference between feedback on the person and feedback on the task? | What should the team member, not you, bring to a one-to-one? |
| Name one sign you’ve delegated the task but kept the decision. | Why does the course suggest giving feedback soon after the event? | How often did the course suggest holding one-to-ones, and why? |
| A team member says the task is “nearly done” for the third week. What do you ask? | Rewrite this into specific feedback: “Great job on the report.” |
Rotate the older questions each time, so module 5’s set drops some module 1 questions and picks up others. That keeps every topic coming back on a regular rhythm, which is the spacing the research supports. You don’t need expanding intervals.
If you’d rather plan this in a spreadsheet, our free Retrieval Practice Kit does it for you. It’s one of the templates in The Evidence-Based Course Design Kit: a Google Sheet where you list the ideas that matter in each module, then mark where learners will recall each one again, such as a quick quiz two modules later or a scenario question at the end. The grid shows at a glance which ideas you only ever test once.
3. Ask for recall, in the format learners will need
Short typed answers, or a prompt to answer in your head before revealing the answer, ask more of memory than picking from a list. But match the format to the end point: the classroom meta-analysis found larger effects when quizzes matched the final assessment. If the course ends with scenario questions, make some review questions scenarios, like the third question in the delegating column. Our guide to Bloom’s taxonomy examples covers why fact-only quizzes don’t prepare learners to apply ideas, and our guide to writing multiple-choice questions covers how to make a choice question test reasoning rather than recognition.
4. Show the answer straight after each question
In a classroom, the teacher marks the grid. In a self-paced course, the course has to. After each answer, show the correct one with a sentence on why, and let learners judge their own attempt. Corrective feedback was linked to a larger benefit in the classroom review, and our entry on effective feedback covers what makes an explanation more useful than a tick.
5. Keep it ungraded and drop the points
The review is for learning, not grading, so make it low-stakes: unlimited attempts, no effect on completion, no score that unlocks anything. Skip the points-for-older-questions scheme. In a classroom it adds a game; on a screen it mostly adds a sum, which is Watson’s working-memory concern. If you want older material to feel worth the effort, say so in one line at the top instead: “The questions from earlier modules are the ones you’re most likely to have forgotten.”
6. Tell learners why it feels hard
Roediger and Karpicke’s students who re-read felt more confident and did worse. Your learners will feel the same pull towards re-watching the video. A sentence at the start of the first review set helps: “You’ll forget some of these. That’s expected, and trying to remember is what makes it stick.”
How the platforms support it
A self-paced retrieval grid needs a few things from a platform: somewhere to put a short practice set inside a module, a way to show the answer after each question, and a way to keep it out of the grade. We haven’t yet published our reviews of Thinkific, LearnWorlds, Teachable and Kajabi, and features change often, so check these in your platform’s current documentation, on the plan you’re paying for:
- Ungraded quizzes anywhere in a module. Can you place a quiz at the start of a module, not only at the end, and set it so it doesn’t count towards completion or a certificate?
- Feedback per question. Can you show an explanation after each answer, rather than only a score at the end?
- Short-answer questions. Is there a free-text question type, and can you show a model answer after learners submit, even if it isn’t auto-marked?
- Question banks across modules. Can a quiz draw questions at random from banks tied to earlier modules, so each review set mixes old and recent material without you rebuilding it each time?
- Unlimited attempts. Can learners retake the set as often as they like?
If your platform can’t mix questions from earlier modules automatically, build each review set by hand. It’s a few minutes per module, and the questions can come straight from the ones you wrote for earlier quizzes.
Further reading
- Roediger & Karpicke (2006), “Test-enhanced learning”. The classic experiment, short and readable, and the clearest demonstration that re-reading feels better than it works.
- Yang et al. (2021), “Testing (quizzing) boosts classroom learning”. The largest review of quizzing in real classrooms, with the moderator analyses on feedback, timing and format.
- Latimier, Peyre & Ramus (2021), “A meta-analytic review of the benefit of spacing out retrieval practice episodes on retention”. The evidence for spacing recall out, and against the need for ever-longer gaps.
- Agarwal (2018), “Download, adapt, and create Retrieval Grids”. The grid as Kate Jones designed it, with links to her blank templates.
- Watson (2020), “Retrieval grids: The good, the bad, and the potential solutions”. A short, fair critique of the grid’s points system and how to fix it.
For the related techniques, see our entry on elaborative interrogation, or browse our research library.
Check your understanding
Three quick questions on this article. Recalling it now helps it stick.
Answer to see the results
Bars show how readers of this page answered. Answers are anonymous.
Questions people ask
What is a retrieval practice grid?
A retrieval practice grid is a table of short recall questions, grouped by when the material was taught: last lesson, last week, two weeks ago, last term. Learners answer from memory, often in any order, and older questions are sometimes worth more points. Teachers use it as a lesson starter. It popularised a simple way to combine retrieval practice with spaced review.
Is the retrieval grid evidence-based?
Its ingredients are. Retrieval practice and spacing both have strong support from experiments and classroom studies. The grid format itself, with its colour-coded rows and points for older material, hasn't been tested in a published study, as far as we can find. So it's a reasonable way to package two well-supported strategies, but its specific features are design choices, not findings.
What is the difference between retrieval practice and the testing effect?
They describe the same thing from two sides. The testing effect is the finding that taking a test improves later memory more than re-studying does. Retrieval practice is the strategy built on it: deliberately asking learners to recall information. Researchers often use the terms interchangeably. The tests don't need to be graded; low-stakes quizzes and free recall both count.
How often should learners do retrieval practice?
More than once, and spread out. In the classroom meta-analysis by Yang and colleagues, more quizzing occasions were linked to larger gains, and a meta-analysis of 29 studies found spacing retrieval out beat doing it all at once. Expanding the gaps each time was no better than keeping them even, so a regular review at the start of each module is a sound default.
Can a retrieval grid use multiple-choice questions?
It can, but free recall asks more of memory, so most grid designers prefer short written answers. The classroom meta-analysis adds a practical rule: quizzes helped more when their format matched the final assessment. If your course ends with a scenario-based multiple-choice exam, practise with that format; if learners will need to explain or write, ask them to do that.
How do you use a retrieval grid in an online course?
Put a short review set at the start of each module with questions from the last lesson, the last module and earlier modules. Ask for short typed answers or a mental answer, then reveal the correct answer straight after each question so learners can check themselves. Keep it ungraded, skip the points tally, and rotate the older questions so the review stays spaced.
Sources
- Roediger, H. L., III, & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249–255. https://doi.org/10.1111/j.1467-9280.2006.01693.x
- Yang, C., Luo, L., Vadillo, M. A., Yu, R., & Shanks, D. R. (2021). Testing (quizzing) boosts classroom learning: A systematic and meta-analytic review. Psychological Bulletin, 147(4), 399–435. https://doi.org/10.1037/bul0000309
- Latimier, A., Peyre, H., & Ramus, F. (2021). A meta-analytic review of the benefit of spacing out retrieval practice episodes on retention. Educational Psychology Review, 33, 959–987. https://doi.org/10.1007/s10648-020-09572-8
- Agarwal, P. K. (2018). Download, adapt, and create Retrieval Grids to boost learning. RetrievalPractice.org. https://www.retrievalpractice.org/strategies/2018/9/28/retrieval-grids
- Watson, A. (2020). Retrieval grids: The good, the bad, and the potential solutions. Learning and the Brain blog. https://www.learningandthebrain.com/blog/?p=5363
The month in learning design
New research on the future of learning, what it means for your course, and a spotlight on subscribers we select at random to feature their products, courses and discoveries. Free, once a month.