Blocked practice and mixed practice
Imagine you are learning four kinds of math problem, which we will call A, B, C and D. The usual way to practice is in blocks: ten problems of type A, then ten of type B, and so on. This is called blocked practice. Interleaving means shuffling them, so that you might meet B, then D, then A, then C, with each type returning several times in the session.
Blocked practice feels productive. After a few problems of the same type you get faster and make fewer errors, which is encouraging. Interleaved practice feels messier. You are slower, you make more mistakes early on, and you may feel that you are not mastering anything. Yet in a number of studies, mixed practice has led to better performance on a later test, particularly when that test mixes problem types too.
Why it may help
The main explanation is that mixing forces you to do something blocked practice lets you skip: decide which method a problem needs. In a block, the method is given away by the previous problem. You apply the same procedure automatically. In a real exam or a real situation, nobody tells you what kind of problem it is. Interleaving rehearses that choice.
A second idea is that interleaving makes it easier to notice differences between similar categories. If you study two types of bird, two painting styles or two similar formulas one after the other in separate blocks, the contrasts can blur. Seeing them side by side makes the distinguishing features stand out.
A third contributor is that mixing naturally spaces out practice on each topic, since a gap lies between one A problem and the next. This overlaps with the benefits of spaced repetition, and the research cannot always separate the two effects cleanly. It also fits the broader notion of desirable difficulties, in which a slower and more effortful session can pay off later.
What the evidence looks like
Interleaving has been tested with mathematics, motor skills, art recognition and science concepts. Many experiments report an advantage for mixed practice on delayed tests, sometimes a large one. But the research is not uniform. Some studies find little or no benefit, some find that the effect depends on how similar the categories are, and many are short lab or classroom experiments rather than whole courses. It is reasonable to describe interleaving as promising and often useful, not as a guaranteed improvement.
When it helps most and when to wait
Interleaving works best when the problems are genuinely confusable and when the learner can already do each type at least a little. Examples include:
- Choosing between different formulas for area, volume or probability.
- Deciding which grammar structure suits a sentence in a language you are learning.
- Identifying which type of logical error or statistical test applies to a situation.
- Practicing different chess tactics, or different musical scales.
It is a poor fit for the very first encounter with a topic. A complete beginner who has not yet seen a single example of each type has nothing to discriminate between and may simply be overwhelmed. Here it is usually better to begin with worked examples and some blocked practice, then mix once the basics are in place. That sequence respects the limits described in the article on cognitive load: you cannot compare methods you have not yet learned.
How to try it
A simple approach needs no special tools.
- Learn each type first. Make sure you can solve at least a few of each type with notes if needed.
- Make a mixed set. From your textbook, past exercises or an open resource, take a few problems of each type. Write them on cards or list them on a sheet, then shuffle.
- Name the method before solving. For each problem, write one line on which approach applies and why. This is the discrimination step.
- Check and note errors. When you pick the wrong method, record what clue you missed. These notes become useful study material.
- Recycle the set. Come back to the same mix a few days later, in a different order.
A short hypothetical example
Suppose a student is revising three kinds of percentage problem: finding a percentage of a number, finding a percentage change, and finding an original value before a discount. In blocked practice, the student does eight of each in turn and feels fluent. In an interleaved set, the problems arrive in random order, and the student pauses on each one to ask whether it is asking for a part, a change or a starting value. At first the student gets more wrong. Over time, the question “which of these is it?” becomes a habit. The student may do better on a mixed test next week, although in practice results vary from person to person.
Common concerns
- “I feel like I’m not learning.” This is expected. Judge by what you can do days later, not how smooth the session felt. If you want a simple check, keep progress tracking notes showing scores on mixed quizzes over time.
- “It’s too hard.” Reduce the number of types mixed at once, perhaps to two, or return to blocks for a short while.
- “My course is organized by topic.” You can still add a mixed review session each week alongside the regular work.
- “Does it apply to facts?” Evidence is less consistent for plain factual recall, where spacing and retrieval practice are the better-supported tools.
The sensible conclusion is modest: once you have the basics of several related skills, mixing them in practice is a low-cost experiment that often improves your ability to choose the right approach, even though it may lower how good you feel during the session.