You have a maths test on the area of rectangles, triangles and circles. You prepare properly: first ten rectangle problems, then ten triangle problems, finally ten circle problems. It all goes well; by the end you're calculating without thinking. Then the test comes, the problems are all mixed up, and suddenly you don't know which formula to use.
You knew the formulas. What you never practised was choosing: recognising which method a problem needs. Interleaving trains exactly that.
What is interleaving?
Interleaving (also called mixed practice) means practising different types of problems mixed together, not in blocks. If there are three types of problem (A, B and C), you don't solve them in the order AAA BBB CCC, but, for example, B A C A C B.
Blocked: three A, three B, three C. Mixed: the same nine problems shuffled: B, A, C, A, C, B, C, A, B.
Blocked practice is the usual way: textbooks go chapter by chapter, and every chapter ends with problems of the same kind. There's nothing wrong with that on its own; you need it to learn something new. The catch is that in blocks you never have to decide which tool you need, because the chapter title has already told you.
Why does it work?
In 2007, Doug Rohrer and Kelli Taylor ran an experiment with university students. They learned to calculate the volume of different solids (like wedges and cones). One group practised in blocks, the other mixed, with exactly the same problems.
During practice the blocked group was far better: they solved 89% of the problems correctly, the mixed group only 60%. But a week later came the test, and the picture flipped. The mixed group scored 63%, the blocked group just 20%.
During practice, the blocked group did better (89% vs 60%), but on the test a week later the mixed group did far better (63% vs 20%).
- blocked
- mixed
What happened? In blocks, students learned to use the formulas, but not which one to use when. Mixed, they had to look at every problem afresh, work out its type and pick the right method. That's slower and harder, and it's exactly what a test asks of you.
It doesn't only work in the lab. In 2015 Rohrer and colleagues tried it in real seventh-grade classes, with homework over several months. On a surprise test a month later, students who had practised mixed scored 72% on average, those who had practised in blocks 38%. In their big 2013 review, Dunlosky and colleagues rated interleaving as a "moderate utility" technique: promising, but with less research behind it than recall or spaced repetition.
There's a second reason, too. When problems come mixed, there are natural gaps between them: while you solve a circle problem, the triangles get a little "rest". That's a form of spaced repetition, which has its own article here.
Your gut feeling lies here too
In 2008, Nate Kornell and Robert Bjork taught people painting styles: participants had to learn to recognise the work of 12 artists. Those who saw the paintings mixed recognised far more later. Yet most participants said blocked learning had worked better. Mixing feels harder, and it's easy to confuse that with worse.
When it works, and when it doesn't
Interleaving isn't a miracle cure, and it's worth knowing its limits. In 2019, Tobias Brunmair and Tobias Richter pooled the results of 59 studies. They found that interleaving helps moderately on average, but it matters a lot what you mix:
- It works best with similar things that are easy to mix up. For example similar types of maths problems, similar painters, similar symptoms of illnesses. The point is noticing the differences.
- It matters less for completely different things. If the problems are very different, there's nothing to confuse.
- It didn't help with learning vocabulary, and in some experiments it even made things worse.
- With brand-new material, start blocked until you understand the basics, and only mix after that.
How to do it
- Learn new material in blocks first, so you understand each method.
- Then mix the problems. Collect problems from several chapters and put them on one sheet, in random order.
- Before each problem, stop for a moment: "What type is this? How can I tell?" That step is the whole point.
- Mix old material with new. Your weekly practice should always include a few problems from earlier chapters.
- Don't worry if it's slower. Making more mistakes than in blocks doesn't mean you're doing it wrong.
Example
Physics, the motion chapter
The chapter has three kinds of problem: constant velocity, constant acceleration and free fall. Instead of doing 8, 8 and 8 problems in textbook order, you write the problem numbers on slips of paper, shuffle them and draw one at a time. For each problem, you first just write in the margin: "acceleration", "free fall", "constant". Only then do you calculate.
Try it: which formula do you need?
The same nine problems in two orders. Blocked first, then mixed. Which one makes you think harder?
1/9
A room is 7 m long and 4 m wide. How many m² of flooring do you need?
Tip
You don't need a special problem book. Collect problems from your old tests and from different chapters of the textbook, and shuffle them. If you don't know which chapter a problem is from, that's perfect: now you have to decide.
What it's good at
- Teaches you to choose between methods, not just to use them.
- Pays off a lot in the long run, especially in maths and science.
- Spaced repetition built in: old material keeps coming back.
- Takes no extra time, just a different order.
Where it falls short
- Practice is slower and you make more mistakes, which can be frustrating.
- With brand-new, hard material you need to understand the basics in blocks first.
- Research says it doesn't help with vocabulary.
- Little benefit when the topics are very different.
Combine it with other techniques
Interleaving goes well with spaced repetition: if every review gives you old and new problems mixed, you're spreading and mixing at the same time. Active recall works here too: the question "what type of problem is this?" is itself a recall.
Mini test: what do you remember?
No scrolling back up. That's the whole point.
Sources
- Rohrer, D., & Taylor, K. (2007). The shuffling of mathematics problems improves learning. Instructional Science, 35(6), 481–498. doi:10.1007/s11251-007-9015-8
- Kornell, N., & Bjork, R. A. (2008). Learning concepts and categories: Is spacing the "enemy of induction"? Psychological Science, 19(6), 585–592. doi:10.1111/j.1467-9280.2008.02127.x
- Brunmair, M., & Richter, T. (2019). Similarity matters: A meta-analysis of interleaved learning and its moderators. Psychological Bulletin, 145(11), 1029–1052. doi:10.1037/bul0000209
- Rohrer, D., Dedrick, R. F., & Stershic, S. (2015). Interleaved practice improves mathematics learning. Journal of Educational Psychology, 107(3), 900–908. doi:10.1037/edu0000001
- Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students' learning with effective learning techniques. Psychological Science in the Public Interest, 14(1), 4–58. doi:10.1177/1529100612453266