Interleaving: why mixing topics when you review works
What interleaved practice is, why mixing topics beats practising one at a time for many subjects, and how to build mixed review sets from your own notes.
Interleaving means mixing different topics or types of problem in one practice session, instead of practising one topic at a time. In maths, for example, you would work through a set where each problem might need a different method, rather than ten problems in a row on the same one. Studies have found that mixed practice often feels harder and goes worse during the session, but leads to better results on a later test, especially when you have to tell similar-looking problems apart. This article explains what the research shows, where interleaving helps and where it doesn't, and how to build mixed review sets from your own notes.
The answer in brief
- Blocked practice: one topic at a time. Interleaved practice: topics mixed together.
- Mixing helps most when you must choose the right method or category, not only apply one.
- Interleaving feels harder and slower during practice; the benefit shows up later.
- Learn a new topic in a block first, then mix it with older topics when you review.
- A review rated interleaving as moderately useful, below practice testing and spacing.
What is interleaved practice?
Most textbooks and courses are organised in blocks. Chapter 5 teaches one method, and its exercises all use that method. Chapter 6 teaches the next. This is blocked practice: AAA BBB CCC.
Interleaved practice mixes them: A C B, B A C, C B A. Each question might come from any of the topics you have covered, and part of the task is working out which method or idea it calls for.
| Blocked practice | Interleaved practice | |
|---|---|---|
| Order | One topic or problem type at a time | Topics mixed within a session |
| How it feels | Smooth; you get faster within the block | Harder; each question makes you think again |
| Performance during practice | Usually better | Usually worse |
| What it trains | Applying a method you already know is needed | Choosing the right method, then applying it |
| Later test results | Often worse | Often better, especially for similar problem types |
What does the research show?
Mixed maths problems
In a study by Rohrer and Taylor (2007), students practised maths problems either grouped by type, as a textbook would present them, or mixed. When they were tested a week later, performance was far better after mixed practice. The authors' conclusion: both of their experiments favoured the shuffled format over the standard one.
Learning to recognise styles
In a study by Kornell and Bjork (2008), people learned to recognise the styles of 12 painters from examples of their work. Seeing paintings by different artists mixed together led to better recognition of new paintings than seeing each artist's paintings in a block. Yet most participants believed that blocking had helped them more.
That mismatch matters. Interleaving is one of the methods that feels worse while it works better, which is why most people don't use it on their own.
How researchers rate it
The 2013 review of study techniques by Dunlosky and colleagues rated interleaved practice as moderately useful. The two highest ratings went to practice testing and spaced practice. The authors noted that the benefits of interleaving had only begun to be explored systematically, so the evidence base is narrower than for those two.
Why does mixing help?
Two explanations are discussed in the research, and both are useful to know as a learner.
It trains you to choose, not only to apply
In a block of identical problems, you know the method before you read the question. Real tests don't tell you. Mixing forces you to look at each problem and decide what kind it is. That decision is often the hardest part of an exam question, and blocked practice never practises it. Later research on the painting task suggests much of the benefit comes from seeing different categories side by side, which helps you notice what distinguishes them.
It builds in effort
Each switch makes you bring a method back from memory rather than reusing the one from the last question. Elizabeth and Robert Bjork describe interleaving, spacing and testing as desirable difficulties: conditions that make practice harder in ways that support long-term learning.
Interleaving and spacing are related but different
Mixing topics also spaces out your practice of each one, since there are other questions in between. But they are different ideas. Spacing is about the gaps between reviews of the same thing, over days. Interleaving is about mixing different things within a session. You can do both, and for review you should. For spacing, see spaced repetition: a simple review schedule.
When does interleaving help, and when doesn't it?
| Situation | Interleave? | Why |
|---|---|---|
| Maths or statistics problems that use different methods | Yes | Choosing the method is part of the skill |
| Similar concepts that are often confused, such as two kinds of statistical test, or two cloud storage services | Yes | Side-by-side practice highlights the differences |
| Recognising categories: painting styles, plant species, types of error message | Yes | Mixed examples teach what distinguishes the categories |
| Exam revision across a whole outline | Yes | The exam itself will mix topics |
| A brand-new topic you don't understand yet | Not yet | Learn it in a block first, then mix |
| Topics with nothing in common | Less clear | There is little to tell apart; spacing still helps |
The practical rule: block to learn, mix to review. When you first meet a method, practise it on its own until you can do it. From then on, mix it with the methods you learned before.
How to build an interleaved review set
You can make mixed practice from any course, using your own questions or the textbook's exercises.
Pick three or four topics you have already learned. They should be related enough that you could confuse them, such as different chapters of the same course.
Collect 4 to 6 questions or problems from each. Your own recall questions, end-of-chapter exercises or practice exam items.
Remove the labels. Copy them into one list without the chapter or topic name next to each.
Shuffle the list. Make sure no more than two in a row come from the same topic.
For each question, first say what kind it is. "This is a two-sample test" or "this is about the shared responsibility model". Then answer it.
Check both: did you identify the type correctly, and did you answer correctly? These are two different skills.
Note every misidentified question. These are confusions between topics. Write a comparison question for each pair: "How do I tell a paired test from an independent one?"
Next time, add the newest topic to the mix and drop one you know well.
A worked example
Say you are studying introductory statistics and have finished the chapters on confidence intervals, hypothesis tests with one sample, tests with two samples, and the chi-square test. Blocked review would mean an evening on each. Mixed review is one sheet with 16 problems in shuffled order, chapter headings removed.
The first problem gives two groups' measurements and asks whether they differ. Before calculating anything, you write "two-sample test". The second asks for a range of plausible values for one group's mean: "confidence interval". The third is a table of counts in categories: "chi-square". You get the type wrong on two problems, both times confusing a paired design with two independent groups. Those two misses go on your list as a comparison question for next time. The calculations were familiar; the choosing was the part you hadn't practised.
What interleaving looks like in different subjects
- Maths and statistics: a problem set drawn from several chapters, with the chapter names removed. Say which method each problem needs before you start it.
- Programming: a mix of small tasks that need different tools: one that needs a loop, one a dictionary, one a function that calls itself. Write each from a blank file.
- Languages: mix tenses, cases or verb groups in one drill, instead of a page of the same ending. Translate sentences that each need a different form.
- Science: mix questions on processes that are often confused, such as two cycles or two kinds of reaction, and explain what makes each one what it is.
- Certification exams: shuffle your questions across all the domains of the outline, as the exam will.
- History or law: mix cases, events or arguments, and ask which principle or cause each one illustrates.
In each case, the move is the same: put things that could be confused next to each other, and make yourself decide which is which before you answer.
Common mistakes with interleaving
- Mixing too early. If you can't yet do any of the problem types on their own, a mixed set is only confusing. Learn each first.
- Mixing unrelated topics and expecting the same benefit. The gain comes mostly from telling similar things apart. A random mix of unrelated facts adds less.
- Judging by how the session felt. Mixed practice feels worse. Judge it by a test a few days later, not by how smoothly the session went.
- Leaving the labels on. If each question shows its chapter, you skip the step of identifying the type, which is the point of mixing.
- Only mixing in practice exams. One practice exam a week isn't enough mixing. Use shorter mixed sets in your regular review too.
Interleaving for exam revision
Exams mix topics by nature. A certification exam draws questions from every domain in its outline, in no particular order. A final exam covers the whole course. Revising one chapter at a time leaves you unpractised at the switching the exam will demand.
As the exam gets closer:
- Shift from topic-by-topic review to mixed sets across the whole syllabus.
- Use mixed practice exams, but treat them as practice, not only as a score.
- Keep a list of confusions between topics and practise them directly.
- Go back to blocked practice only for a topic you can't do at all, then put it back into the mix.
For planning revision from an exam's official outline, see how to study for a certification exam from its outline. For keeping track of which confusions keep happening, see metacognition for students: find out what you don't know.
Doing it in Soundings
Soundings mixes your questions in a few places. Quiz yourself, on the Questions board, asks your own answered questions in a shuffled order. On the Questions page, choose Every course to mix questions from all your courses, or one course to mix its topics (Quiz yourself and use recall mode). You can quiz on all answered questions, on the ones you got wrong before, or on the ones not yet mastered, and a quiz doesn't change when your recall cards come back. In Review, Quiz yourself instead starts a quiz on the course's questions or all courses' (Review your cards).
For learning a topic in a block first, open it on the course's Topics tab and use recall mode, which covers each answer on the topic's page (Organise with topics). When you get a quiz question wrong, Soundings asks what you missed, and Mixed it up is one of the reasons you can record.
Common questions
What is the difference between interleaving and spacing?
Spacing means leaving gaps of days or weeks between reviews of the same material. Interleaving means mixing different topics or problem types within one session. They often go together, because mixing also puts gaps between questions on the same topic, but they are separate ideas.
Why does interleaving feel harder?
Because each question makes you work out what kind of problem it is and bring back the right method, instead of repeating the method you used a moment ago. That extra effort is part of why it works. Studies have found people often believe blocked practice helped them more, even when mixed practice gave better results.
Should I interleave when learning something new?
Usually not at the very start. Learn a new method or concept in a block until you can do it, then mix it with earlier topics when you review. Interleaving helps you tell things apart, which needs you to know each thing first.
Does interleaving work for all subjects?
The clearest evidence is for maths problems and for learning to recognise categories, such as painting styles. It is likely to help wherever you must choose between similar methods or concepts. For topics with nothing in common, the benefit is less clear, though spacing still helps.
How many topics should I mix?
Three or four related topics is a good start, with several questions from each, shuffled so the same topic rarely comes up twice in a row. As your course goes on, keep adding the newest topic and dropping one you know well.