Interleaving: when mixing topics helps you learn, and when it does not

Interleaving is a study method in which you mix several related topics or problem types in one session. You do not finish one topic before you start the next. In the notation of Taylor and Rohrer (2010), the two orders are abcbcacab (interleaved) and aaabbbccc (blocked). Interleaving helps most when the things you study are similar and easy to confuse. Every switch makes you look for the feature that tells them apart. It does not help with everything. In a meta-analysis of 59 studies, the average benefit was moderate, and studies that used words as the learning material favoured blocked study.

• Interleaved practice means mixed practice: A, B, C, B, A, C. Blocked practice means one topic at a time: A, A, A, then B, B, B.
• The best-known studies found large effects. In one, students solved 63% of new maths problems a week after mixed practice, against 20% after blocked practice.
• Across many studies, the effect depends on the material. It was largest for paintings and smallest, or reversed, for word material.
• Interleaving feels worse while it works. Practice scores drop. Most learners in the studies judged blocked study to be at least as effective, even when their own results showed the opposite.
• For psychology students, the evidence points to a split. Memorise each definition with spaced retrieval, then mix concepts that are easy to confuse and practise telling them apart.

What interleaving means

Interleaving is a way to order practice. It does not add study time or change the material. Take the four schedules of reinforcement from operant conditioning: fixed ratio, variable ratio, fixed interval and variable interval. In blocked practice, you answer ten questions about fixed ratio, then ten about variable ratio, and so on. In interleaved practice, the same 40 questions come in a mixed order, so no two questions in a row are about the same schedule.

The two orders train different things. In a block, you know the answer type before you read the question, so you only need to apply one rule. In a mixed set, you must first decide which concept the question is about. That decision is the part an exam tests, and blocked practice skips it. Our explainer on the schedules of reinforcement shows why these four are easy to mix up: they differ on two features.

Research on motor skills gives a classic example. Robert and Elizabeth Bjork describe a 1979 study in which people practised three movement patterns, either in blocks or in a random order. The blocked group improved faster during training. Ten days later, under a random test order, the interleaved group performed far better. The Bjorks list interleaving among the desirable difficulties: conditions that make learning feel harder but lead to more durable and flexible learning.

The evidence: three classic studies

The numbers that study guides quote come from three lines of research, each with a different kind of material.

The first is maths. In a study by Rohrer and Taylor (2007), college students learned to calculate the volume of four unusual geometric solids. During practice, the blocked group got 89% of the problems right and the mixed group only 60%. One week later, the result reversed: the mixed group solved 63% of the test problems and the blocked group 20%. Only 18 students completed all three sessions, and mixing also put more time between problems of the same type. A later study with children fixed the amount of spacing, and interleaving still doubled the scores on a test one day later.

The second is art. Kornell and Bjork (2008) showed participants six paintings by each of 12 artists. For six of the artists, the paintings came one after another. For the other six, they were mixed with paintings by other artists. On new paintings, participants picked the correct artist 61% of the time for the mixed artists and 35% for the blocked ones. An exact, high-powered replication of their second experiment found a medium-sized advantage, similar to the original.

The third is the classroom. In a preregistered trial by Rohrer and colleagues (2020), 54 seventh-grade maths classes did interleaved or blocked assignments for four months. On an unannounced test one month later, the interleaved classes scored 61% and the blocked classes 38%. The authors add that “important caveats remain”. The effect also appears at university level. Physics students who did interleaved homework solved more problems on two surprise tests, with median improvements of 50% and 125%.

What a meta-analysis of 59 studies found

Single studies show that interleaving can work. A meta-analysis shows how often and for what. Brunmair and Richter (2019) combined 238 effect sizes from 59 studies. The average effect was moderate, but it changed a lot with the learning material. The effect size here is Hedges’ g: positive values favour interleaving, and negative values favour blocking.

The meta-analysis also found three conditions that made the effect stronger. Interleaving helped more when the categories were similar to each other and when the examples inside each category varied more. It also helped more with complex material. The authors conclude that interleaving “can effectively foster inductive learning”. They add that it “should be used with caution in certain conditions, especially for expository texts and words”.

The words result needs care. It rests on only 13 effect sizes, from tasks such as names in different categories, pronunciation rules and translations. The authors’ manuscript warns that the negative effect “might not be caused by the use of words but rather by the task types”. A systematic review by Firth, Rivers and Boyle (2021) covered 26 studies. It found interleaving “of greatest use when differences between items are subtle”. It also noted that most of the evidence comes from laboratory studies with university undergraduates.

When interleaving does not help

The limits matter as much as the headline numbers, because most study guides leave them out. Three situations stand out in the research.

The first is mixing topics that have nothing in common. Many students study two subjects from flashcards and keep the two decks apart. Hausman and Kornell (2014) tested whether mixing the decks helps. Participants either alternated between anatomy terms and Indonesian vocabulary, or studied one topic and then the other. In four experiments, mixing made no reliable difference. The paper’s title says it plainly: “Mixing topics while studying does not enhance learning.”

The second is categories that are easy to tell apart. Carvalho and Goldstone (2014) found that interleaving gave better results for categories with high similarity. For categories with low similarity, blocked study was better. Their explanation is that interleaving shows the differences between categories, and blocking shows what the members of one category have in common.

The third is word material, as the meta-analysis showed. When you learn what a term means, the task is close to learning a word pair: term on one side, definition on the other. This is our reading, not a tested result. The meta-analysis found no benefit for word material, and the few word studies favoured blocking. The flashcard study above also found no benefit from mixing vocabulary with anatomy terms.

A major review of study techniques by Dunlosky and colleagues (2013) also rated interleaved practice as “moderate utility”. Only practice testing and distributed practice got the high rating. The authors wrote that the benefits of interleaving “have just begun to be systematically explored”.

Is interleaving just spacing in disguise?

No, although the two are connected. When you interleave, repetitions of each topic are automatically spread out, and spaced repetition improves memory by itself. So the early results could have come from spacing, not from mixing.

Two studies tested this directly. In the maths study with children, the researchers kept the spacing equal in both orders, and interleaving still doubled the test scores. Kang and Pashler (2012) used paintings. When they added time between paintings by the same artist, without mixing artists, the extra time did not help. When they showed paintings by different artists side by side, learners did as well as with interleaving. The authors call the mechanism discriminative contrast: interleaving works because it puts different categories next to each other, so the learner compares them.

In practice, spacing and interleaving answer different questions. Spacing decides when you come back to a topic. Interleaving decides what sits next to it when you do.

Why interleaving feels worse while it works

Learners rate interleaving lower than it performs, and several studies measured this. In Kornell and Bjork’s painting study, 78% of the participants did better with the mixed artists. But 78% also said that blocked study was as good as or better than mixing.

The same pattern appeared with psychology material. Zulkiply and Burt (2013) gave 40 students short case descriptions of six psychological disorders, each with an invented label. The disorders included obsessive-compulsive disorder, phobia and depression. After interleaved study, the students classified new cases more accurately, both soon after study and one week later. Even so, 65% of the short-delay group and 70% of the one-week group judged blocked study to be more effective. In the university physics study, students rated interleaved homework as harder and wrongly believed they had learned less from it.

The Bjorks explain the illusion this way. Blocked practice produces good performance during the session, and learners take that performance as proof of learning. Interleaving lowers practice scores, as the 89% against 60% in the maths study shows, so it feels like failure. Only a delayed test shows which method produced more learning.

How to use interleaving for psychology terms

Psychology exams ask two kinds of questions. Some ask what a term means. Others describe a case and ask which concept it shows. The research supports a different method for each, in this order.

1. 1.Learn each definition first. Use retrieval and spacing for this stage, not mixing. Our guide to memorising psychology terms explains why practice testing and spacing have the strongest evidence.
2. 2.Build small sets of concepts that are easy to confuse. The studies in this article do not give an ideal set size, but three to six terms that share features are easy to build and check. Do not group terms that have nothing in common.
3. 3.Practise with cases, not only definitions. Read a short description, decide which concept it shows, then check. This is the task where interleaving had its benefit, as in the disorder case study.
4. 4.Start with blocks if a set is new, then mix. In a painting study by Yan and colleagues (2017), a blocked start followed by mixing, and short mini-blocks, worked statistically as well as pure interleaving.
5. 5.Expect lower scores in practice. A drop in practice scores is not a sign that the method fails.
6. 6.Come back to each set on later days, so you get spacing and contrast together.

Good sets come from the pairs that students mix up most. Searches that reach this site show some of them: people look up “retroactive interference vs retrograde amnesia”, two terms from different parts of memory research.

Practising mixed sets in Psychology World

In Psychology World you can save terms like these to one list, one list for each set of easily confused concepts. Then you can drill the list as flashcards, as multiple-choice questions once it has four or more terms, or by typing the answer. The four reinforcement schedules, or the four operant consequences, meet the four-term minimum for multiple-choice questions by themselves. What each study mode does is set out on the home page.