Interleaving
Also known as mixed practice, varied practice
Mixing different kinds of problem in one session beats doing all of one kind, then all of the next.
The claim Practising several related skills in a shuffled order produces worse performance during practice and better performance on a later test than practising each one in its own block.
Good evidence Replicated across motor skills, categories and mathematics, including a preregistered classroom randomised trial; the size of the benefit outside maths is less well mapped.
Almost every textbook, worksheet and course is blocked: ten problems on one technique, then ten on the next. Interleaving shuffles them, so no two consecutive problems use the same method. The immediate consequence is that practice gets harder and messier, accuracy during the session drops, and it feels like the wrong thing to be doing.
The strongest evidence for it is unusually good for this field, because it is a real classroom trial rather than a lab session. Rohrer and colleagues ran a preregistered, cluster-randomised study across 54 seventh-grade mathematics classes for a full term. On an unannounced test a month later, the interleaved group scored 61% and the blocked group 38%. An earlier study by the same group found the same pattern with a smaller sample: after thirty days, 74% against 42%.
The explanation people find most convincing is about choosing, not about remembering. With blocked practice you know the method before you read the question, because it is the method from the last nine questions. What you never practise is the step that a real test demands first: working out which method this problem calls for. Interleaving forces that discrimination on every item.
There is a second mechanism worth separating out. Interleaving smuggles in spacing — if problems of a type are scattered across an assignment rather than clumped, the gaps between them are longer. Some of the benefit is therefore the spacing effect in disguise, and the studies that control for this find an interleaving effect remaining on top of it.
The condition for it working is that the things being mixed are confusable. Interleaving is a discrimination trainer. Shuffling topics that could never be mistaken for one another adds difficulty and buys nothing, which is the most common way people try it and conclude it does not work.
What to do differently
- Mix problem types within an assignment rather than finishing one type before starting the next.
- Interleave things that are genuinely easy to confuse — that is what the technique trains. Unrelated subjects do not qualify.
- Accept a lower hit rate during practice. If your accuracy while practising did not fall, you probably have not interleaved.
- When making your own practice set, strip the labels: a problem that announces its own method has already given away the part worth practising.
- Start blocked and then interleave. A method you cannot yet execute at all is not ready to be discriminated from another.
Limits and trade-offs
- It reliably makes practice performance worse, and both learners and teachers read that as evidence it is failing — the effect is invisible until a delayed test.
- Beginners need some blocked practice first. You cannot choose between two techniques you cannot yet carry out.
- The classroom evidence is concentrated in mathematics. Extending it to essay writing, programming or physical skills is a reasonable bet, not a demonstrated result.
- Interleaving unrelated material is a common misreading and does nothing useful; the benefit depends on the items being confusable.
What this means for a language
The language-learning equivalent of blocked practice is the themed unit — a chapter of food vocabulary, a chapter of travel vocabulary, a drill on one tense at a time — and it has exactly the weakness the maths studies describe: you never practise deciding which tense the situation calls for, because the exercise already told you. Mixing tenses in one drill, or reviewing vocabulary from six unrelated topics together, is the direct translation. Note the limit, though: this literature is about choosing between confusable procedures, and much of language learning is not procedural at all. It has little to say about how much comprehensible input you should get.
The research
- Rohrer, Dedrick, Hartwig & Cheung (2020), Journal of Educational Psychology
Preregistered cluster-randomised trial, 54 seventh-grade classes, four months; unannounced test a month later gave 61% interleaved vs 38% blocked (d = 0.83).
- Rohrer, Dedrick & Stershic (2015), Journal of Educational Psychology
126 seventh-graders; interleaved beat blocked on a test one day later (80% vs 64%) and by more after thirty days (74% vs 42%).
- Dunlosky et al. (2013), Psychological Science in the Public Interest
Rated interleaved practice 'moderate utility' — promising, but with less evidence across materials and ages than testing or spacing.
Related
Read next
-
The single most useful document here. Ten study techniques rated for utility across materials, learners, tasks and settings. Only practice testing and distributed practice earn 'high'; highlighting, rereading and summarising come out low.
-
Preregistered, cluster-randomised, 54 seventh-grade classes, a full school term. On an unannounced test a month later the interleaved group scored 61% against 38%. Rare in this field: a real classroom trial, not a lab session.
-
The standard first book. Two of the three authors are the cognitive psychologists behind much of the retrieval-practice literature, so the claims are first-hand rather than second-hand. Heavy on anecdote between the findings, which some readers find padding.
-
Popular-press, and it argues a thesis rather than surveying a field — but its chapters on interleaving, desirable difficulties and the limits of early specialisation are the most readable version of that literature anywhere.
-
Free, Creative-Commons posters and one-page guides for spacing, retrieval practice, elaboration, interleaving, concrete examples and dual coding. The fastest way to put the evidence on a wall where you will see it.
-
The shortest honest route into the evidence, illustrated throughout. Written by the researchers behind The Learning Scientists, so it doubles as the paper version of their free posters.
-
A cognitive scientist and a classroom teacher writing together, which is why it is unusually specific about what the strategies cost in class time. Aimed at teachers; a self-directed learner has to translate.
-
The review that explains why good technique is not enough: learners systematically misjudge their own learning, and the methods that feel best while studying are usually the ones that leave least behind.