What the research says

Mixing Topics in Practice Works Better

Most maths homework practises one skill at a time. A three-month study with 126 seventh graders mixed problem types instead, and a month later scored far higher on a surprise test.

By the Giga Quizzes team4 min read
A heap of round sweets in mixed colours
Photo: Behnam Norouzi, Unsplash, Unsplash License, cropped.
In this post
  1. 1Blocked and interleaved practice
  2. 2The study
  3. 3What they found
  4. 4Why mixing helps
  5. 5The limits
  6. 6Interleaving in a quiz

Open almost any maths textbook and the practice looks the same: a lesson on one method, followed by a dozen problems that all use that method. It is efficient and it feels productive, because by the tenth problem students are fast and accurate. Doug Rohrer, Robert Dedrick and Sandra Stershic pointed out the catch. If every problem on the page uses the method from today’s lesson, students never have to decide which method a problem needs. On a test, and in the next course, deciding is half the work.

Blocked and interleaved practice

Practice in which every problem uses the same strategy is called blocked practice. The alternative is interleaved practice: problems of different kinds are mixed, so that each one has to be read and recognised before it can be solved.

BlockedInterleaved
Ten slope problems, then ten graphing problemsSlope, graphing, slope, a word problem, graphing…
The method is known before the problem is readThe method must be chosen from the problem itself
Feels smooth and successfulFeels harder and slower
Like the textbookLike the exam

Interleaving does not add new problems or extra time. It rearranges the ones students would have done anyway.

The authors give an example of the usual approach: a dozen problems after a lesson on the Pythagorean theorem, every one of them solved with the Pythagorean theorem. A student can complete the whole set without ever asking whether the theorem applies, because the page has already answered that. Then, on a comprehensive exam or in the next year’s course, the same kind of triangle turns up among problems of every other kind, with no heading to say which method it needs. That is the moment blocked practice never rehearsed.

The study

The experiment ran for three months in seventh-grade maths classes, with 126 students. Every student received the same practice problems over that period. The only difference was the order: the problems were arranged so that skills were learned either through interleaved practice, with part of each assignment mixing different kinds of problems, or through the usual blocked approach.

The practice phase ended with a review session. Then, either one day or thirty days later, students took a test they had not been told about.

What they found

Interleaved practice produced higher scores on both tests. Measured as effect sizes, the advantage was d = 0.42 on the test a day later and d = 0.79 on the test a month later.

The second number is the remarkable one. The advantage was larger after a month than after a day: whatever students gained from mixed practice, it lasted better than what they gained from blocked practice. In education research, where most effects shrink as time passes, that is unusual. An effect of 0.79 is large by the standards of classroom studies, and it came from nothing more than rearranging problems that were already in the course.

Why mixing helps

The authors’ main explanation is in the setup. In a blocked assignment, the hardest decision has been made for the student before they begin: today’s problems use today’s method. Interleaving takes that help away. Students have to look at each problem and ask what kind of problem it is, which is precisely the skill an exam or a new situation demands.

A second reason is that interleaving naturally spaces practice. When the same kind of problem appears on several different days, mixed with others, students meet it again after a gap, which is itself known to improve memory.

Researchers are still working out how much each reason contributes. The first idea, learning to tell kinds of problem apart, also appears outside maths: in a well-known study, people learned to recognise painters’ styles better when the paintings were mixed rather than grouped by artist. See mixed examples teach recognition.

The limits

The study took place in one setting, and the details of how much of each assignment was mixed are particular to it. It mixed problems into practice after skills had been taught; it did not test mixing from the very first encounter with a new method, when a few problems in a row may help students see how the method works. And mixed practice feels harder, so students may judge it as less effective even when it is working. That is worth explaining to them in advance.

Interleaving in a quiz

A quiz is an easy place to interleave, because the order of questions is entirely up to the person writing it.

A mixed review quiz

  • Teach and practise a new method in a short block first, so students understand it.
  • In review quizzes, mix questions from the current topic with questions from earlier ones.
  • Include problems that look similar but need different methods, so students must choose.
  • Shuffle the order rather than grouping by topic.
  • Warn students that mixed practice will feel harder, and that this is the point.

Interleaving pairs naturally with spacing, because mixing old topics into new practice brings them back after a gap. The post on spacing: three short sessions beat one long one covers that side of it.

Sources

  1. Rohrer, D., Dedrick, R. F. & Stershic, S. (2015). Interleaved practice improves mathematics learning. Journal of Educational Psychology, 107(3), 900–908.

More from the blog

All posts →