Interleaving: Why Mixed Practice Feels Worse and Works Better

Practise one skill until it is smooth, then the next: it feels efficient, produces confident practice sessions, and is measurably the weaker way to learn. The alternative — mixing related problem types within practice — feels clumsy and outperforms it where it matters: later, on the test, in the world.

The evidence

Rohrer and Taylor (2007) taught students mathematics with practice either blocked by type or interleaved across types. Blocked practisers looked better during practice; interleaved practisers scored dramatically higher a week later. The pattern has replicated across mathematics topics, category learning and motor skills, and Rohrer’s subsequent reviews (2012) locate the mechanism: interleaving forces the learner to choose a strategy for each problem, not merely execute the one the heading announced. Discrimination — which kind of problem is this? — is most of real competence, and blocked practice never rehearses it.

Why nobody does it voluntarily

Interleaving is a desirable difficulty in Bjork’s sense: the very stumbling that improves retention makes practice feel unproductive. Learners judge learning by current fluency, so they prefer blocking; textbooks oblige with chapters of same-type exercises; and the strongest students, fluent fastest, get the least discrimination practice of all — then meet mixed-topic examinations that test exactly the skill blocking never built.

Using it well

Interleave related material, not random noise: mixing fraction, ratio and percentage problems teaches their boundaries; mixing algebra with French vocabulary teaches nothing. Introduce a new topic with a short blocked burst, then fold it into the mix. Warn learners the stumble is the treatment working, or they will retreat to comfort. And build mixed retrieval into every week rather than saving mixture for mock season.

Our practice sets interleave by design — the discomfort is doing the teaching: the Talent Framework.

References

Rohrer, D., & Taylor, K. (2007). The shuffling of mathematics problems improves learning. Instructional Science, 35(6), 481–498.
Rohrer, D. (2012). Interleaving helps students distinguish among similar concepts. Educational Psychology Review, 24(3), 355–367.