Retrieval Practice and Spaced Practice: The Evidence Behind the Two Strongest Memory Tools

Retrieval practice and spaced practice are the closest thing classroom cognitive science has to a free lunch. They are cheap, compatible with almost any curriculum, and supported by a deep research base going back more than a century. They are also often taught in a way that loses most of the effect — trivial recall, same-day testing, no feedback, no variation. This evidence review sets out what the research actually found, where the classroom evidence is strong and where it is still thin, and how to run retrieval and spacing routines that produce the gains the meta-analyses report.

What retrieval practice and spaced practice are

Retrieval practice is the act of bringing information to mind from memory — not rereading, not re-highlighting, but actively recalling. Spaced practice distributes those retrievals, or restudy episodes, over time rather than massing them into a single block.

The two ideas combine. A single short quiz is retrieval. The same short quiz spread over six weeks with mixed content is retrieval plus spacing. Both effects stack.

Where the research comes from

The basic spacing effect dates to Hermann Ebbinghaus's memory experiments in the 1880s — older than almost every other method in this library. The modern "testing effect" literature, associated with Henry Roediger, Jeffrey Karpicke, Pooja Agarwal and others, has accumulated decades of laboratory evidence that retrieval produces stronger long-term retention than restudy. Classroom translation picked up in the 2010s, with educator-facing work from Agarwal, the Learning Scientists, and later guidance from the Education Endowment Foundation.

What the research actually shows

The headline effects are unusually strong for classroom cognitive science.

The core practice-testing meta-analysis found that testing outperformed restudy and other non-testing conditions. The EEF's summary of the same evidence reports classroom effects around g ≈ 0.67 and laboratory effects around g ≈ 0.62. These are large effects — genuinely meaningful shifts in learning outcomes for a near-zero-cost intervention.

Transfer effects in the retrieval literature are also non-trivial. Adesope and colleagues' summary reports transfer effects of about 0.53 and retention effects of about 0.63. In plain terms: quizzing doesn't just help you remember the exact items you quizzed; it also helps you apply what you know to related problems. A real classroom study in sixth-grade social studies found that quizzing improved chapter and semester-exam performance.

Classroom spacing evidence is newer and somewhat more conditional. Mawson and Kang's 2025 classroom-focused meta-analysis found d ≈ 0.54 for distributed over massed practice, with very high heterogeneity. The stronger effects tended to occur at longer retention intervals and in higher education; primary-school studies were fewer and less consistently positive. A large nine-course STEM study found an overall positive effect only in some analytic choices and concluded that generalisability across subjects was still unclear.

The defensible synthesis: retrieval is one of the most reliable classroom cognitive-science findings we have. Spacing in the lab is equally well-replicated; spacing in classrooms is positive on average but still accumulating evidence, especially in primary education.

Why these effects exist

Retrieval practice strengthens memory because every successful recall is itself a learning event. When a student struggles to pull a definition out of memory and then succeeds, the memory trace is reshaped and reinforced. Restudy just exposes the student to the same information again; retrieval requires the mind to do the work.

Spacing produces a "desirable difficulty." When a learner returns to a topic after enough forgetting has set in, the retrieval is slightly effortful. That effortful retrieval is exactly the condition that produces durable memory. Massed practice feels productive because performance looks good in the moment. Spaced practice feels harder because some items feel forgotten — and that difficulty is exactly where the long-term learning is happening.

Feedback matters. Retrieval without feedback can reinforce errors. A short correction after a wrong answer is what converts the retrieval from an error-entrenching experience into a learning one.

How to run retrieval practice in class

A disciplined retrieval routine has six moves.

• Make sure initial teaching is secure. Retrieval is a memory tool, not a first-learning tool. Quizzing material that was never properly taught creates frustration and false negatives.
• Build cumulative low-stakes quizzes into lessons. A two-minute quiz at the start of every lesson, mixing this week's and previous units', is the canonical format.
• Mix recent and older content. The mix is what creates interleaving and spacing effects simultaneously.
• Prefer short-answer or cued recall. Multiple choice is easier to administer but weaker as a retrieval trigger than short-answer. Use multiple choice where you must; use short-answer where you can.
• Give concise feedback. One sentence of correction is enough. Do not lecture.
• Revisit material over days and weeks. Not all on the same day.

How to run spaced practice in class

The simplest spacing routine is curriculum-mapped cumulative quizzing: each week's do-now includes items from this week, last week, the unit before, and the term before. Students re-encounter every key idea multiple times, at increasing intervals.

A more deliberate approach builds a spacing schedule directly into the unit plan: initial teaching in week 1, revisit in week 2, revisit in week 4, revisit in week 8, revisit at the start of the following term. The intervals roughly follow the expanding spacing schedule the lab evidence supports.

Classroom examples across phases

Primary. Year 4 science. Every Monday morning, a two-minute cumulative quiz on states of matter. Week 1: today's content only. Week 2: this week + last week. By week 6, every quiz covers all the content taught in the unit, mixed randomly. The teacher reads answers in 30 seconds with short corrections.

Secondary. Year 11 history. Every lesson opens with a four-item retrieval starter. One item from today's previous work, one from last week, one from the unit before, one from the previous term. Feedback is a 60-second class discussion.

Tertiary. First-year anatomy. Weekly cumulative low-stakes quizzes on the week's material plus random items from previous weeks. Distributed over a 12-week term, every concept is revisited at intervals of roughly one, three, and eight weeks. Paired with spaced revisit sets in the revision period.

Where retrieval and spacing go wrong

The common failure modes are avoidable.

• Retrieval before learning is secure. Quizzing something students were never properly taught produces frustration, not learning. Teach first, retrieve later.
• No feedback. Wrong answers without correction reinforce the error. Feedback doesn't have to be long; it does have to exist.
• Surface-level items only. If every question is a definition, the retrieval builds rote knowledge and not much else. Mix in application, inference, and comparison items.
• Same-day only. Quizzing the day you taught something is fine. Quizzing only on the day you taught something is not spacing. The benefit grows with the gap.
• Too long, too graded. Retrieval works best low-stakes and brief. Turning every quiz into a summative assessment tanks the routine by increasing anxiety and marking load.

Best fit and limits

Retrieval practice is best fit from later primary onward, in any subject with knowledge worth remembering: vocabulary, factual schemas, equations, definitions, historical knowledge, anatomy, and any foundational content later learning depends on. Spacing fits everywhere retrieval fits, with the evidence stronger at longer retention intervals and in higher-education settings.

The limits are worth naming. The classroom spacing literature is still smaller than the lab literature, especially in primary education. The retrieval effect itself is well-replicated, but genuine classroom spacing studies remain much rarer than laboratory studies, particularly outside mathematics and STEM.

Teacher requirements, assessment, and resources

Resource-light. A spreadsheet or a curriculum map that tracks which material has been taught and when. The investment is curriculum sequencing, a small bank of diagnostic items per unit, and a habit of revisiting content over weeks rather than in a block.

Evaluate through delayed tests rather than same-day scores. Same-day quiz performance tells you who was listening today. A delayed cumulative check at the end of the unit tells you what actually stuck.

For a classroom-ready retrieval template, see Retrieval Practice Lesson Plan.

How TAyumira supports retrieval and spacing

TAyumira supports retrieval practice as one of its ten research-backed teaching methods. For units and courses generated in TAyumira you get:

• Cumulative do-now starters with items drawn from previous lessons, units, and terms
• A built-in spacing schedule at unit level: initial teach, week 2, week 4, week 8 revisit
• Short-answer and cued-recall items with expected correct answers and common misconceptions
• A delayed cumulative check at the end of each unit
• Exports of quiz banks to presenter, worksheet, or digital forms

Start for free — the Free tier covers the full workflow.

FAQ

What is the effect size of retrieval practice?

The EEF's summary of the core practice-testing meta-analysis reports classroom effects around g ≈ 0.67 and laboratory effects around g ≈ 0.62. Adesope and colleagues' summary reports transfer effects of about 0.53 and retention effects of about 0.63. These are large effects for a near-zero-cost intervention.

Is spaced practice as well-evidenced as retrieval?

In the laboratory, yes — the spacing effect is one of the oldest findings in cognitive psychology. In classrooms, the evidence is newer and more conditional. Mawson and Kang's 2025 classroom-focused meta-analysis found d ≈ 0.54 for distributed over massed practice, with high heterogeneity and stronger effects at longer retention intervals and in higher education.

Should I use multiple choice or short-answer for retrieval?

Short-answer where you can, multiple choice where you must. Short-answer requires a genuine retrieval effort; multiple choice only requires recognition. If you use multiple choice, make distractors that map to specific misconceptions so the question still tests understanding.

Do I need to grade retrieval quizzes?

No. The research evidence consistently favours low-stakes, brief retrieval with feedback over high-stakes graded quizzes. Low stakes protects the learning function of retrieval from the anxiety and overhead of summative assessment.

How often should I use retrieval practice?

A two-minute cumulative retrieval starter every lesson is a reasonable baseline. Once a week is enough to build durable memory over a unit; every lesson is stronger. The exact cadence matters less than consistency.

Related evidence reviews

• Explicit Instruction Evidence
• Formative Assessment Evidence
• Metacognition and Self-Regulated Learning Evidence
• Mastery Learning Evidence

Sources

• Adesope, O. O., Trevisan, D. A., & Sundararajan, N. (2017). Rethinking the Use of Tests: A Meta-Analysis of Practice Testing.
• EEF. Does research on retrieval practice translate into classroom practice?
• Mawson, D., & Kang, S. H. K. (2025). Classroom spacing meta-analysis.
• McDaniel, M. A., Agarwal, P. K., Huelser, B. J., McDermott, K. B., & Roediger, H. L. (2011). Test-enhanced learning in a middle school science classroom.

Try retrieval and spacing on your next unit

Write a six-week spacing schedule for the unit you are about to teach. Each week's do-now includes one item from this week, one from last week, one from the unit before. Keep it brief. Feedback in under a minute. If you want the quiz items generated for you in a cumulative schedule, create a free TAyumira account.