How Interleaving Practice Builds Deeper and More Flexible Skills

Athletes Who Varied Their Drills Outperformed Specialists on Match Day

In a 1978 study of baseball batting, researchers divided players into two groups. One group practiced hitting fastballs in blocks — 45 consecutive fastballs, then 45 curveballs, then 45 change-ups. The other group faced pitches in a random, interleaved order. During practice, the blocked group performed noticeably better. In a final test that mixed all pitch types, the interleaved group outperformed the blocked group by a significant margin. The pattern has since been replicated in mathematics, medicine, art history, and surgery. Blocked practice feels productive. Interleaved practice produces skill.

Interleaving — sometimes called mixed practice or varied practice — is a learning schedule in which different problem types, skills, or topics are practiced in an alternating or random order rather than grouped by category. The difficulty it creates is intentional and, according to research, is precisely why it works.

Why Blocking Feels Better but Fails Later

Blocked practice — completing all exercises of one type before moving to the next — produces rapid within-session improvement. Each problem resembles the one before it, making solution selection fast and accurate. Learners leave a blocked session feeling competent. The problem is that this within-session fluency does not reflect the neural processes that build lasting skill.

When all practice items are identical, learners can solve them by holding the solution method in working memory rather than retrieving it from long-term memory each time. The procedure is never truly consolidated because it never has to be rediscovered. The moment the block ends and the context changes, performance collapses.

The Discrimination Hypothesis

Cognitive scientist Doug Rohrer at the University of South Florida, who has conducted extensive classroom research on interleaving, proposes that the primary mechanism is discrimination learning. When problems are interleaved, each new problem requires the learner to first identify which procedure applies before applying it. In real exams and real-world tasks, problems never arrive pre-labeled by type. Interleaved practice rehearses the identification step that blocked practice completely bypasses.

A second mechanism, consistent with Robert Bjork's desirable difficulty framework, is that the added difficulty of interleaving forces effortful retrieval of solution procedures from long-term memory on every trial, producing the same consolidation benefits seen in spaced retrieval practice.

Evidence Across Domains

The Subjective Trap

One of the most striking findings in this research area is the mismatch between perceived and actual learning. Studies consistently find that learners prefer blocked practice, rate it as more effective, and predict they will perform better having used it — even after experiencing evidence to the contrary. Kornell and Bjork (2008) found that after completing an interleaved study session that produced better performance, most participants still reported that blocked practice had worked better for them.

This metacognitive illusion has direct practical consequences. Students who can choose their own practice schedules — the vast majority — systematically choose the method that feels productive rather than the one that is productive. Teachers who rely on student satisfaction ratings may unknowingly be rewarded for using less effective instructional designs.

Implementing Interleaving in Practice

• Mathematics homework: instead of completing an entire chapter's worth of problems on one technique, create mixed problem sets drawing from several recent chapters
• Language learning: alternate grammar rules, vocabulary retrieval, and listening comprehension within a single session rather than dedicating separate sessions to each
• Music practice: cycle through pieces or technical exercises in short blocks rather than drilling one piece to completion before starting the next
• Medical case study review: mix cases from different diagnostic categories rather than studying all cardiac cases, then all pulmonary cases

The degree of interleaving matters. Research suggests that highly random interleaving produces the strongest long-term benefits but also the greatest short-term frustration, which can undermine motivation in novice learners. A practical middle ground for beginners is near transfer interleaving: alternating related but distinct problem types rather than randomly sampling from an entire curriculum.

Interleaving vs. Spacing: Complementary Mechanisms

Interleaving at Scale: Implications for Curriculum Design

Interleaving challenges the standard textbook architecture, in which each chapter covers a single topic exhaustively before moving to the next. Research by Rohrer and colleagues in authentic middle-school classrooms found that textbook chapters redesigned to mix problem types from recent and current units produced significantly better performance on end-of-year tests with no increase in total instruction time.

Some publishers have begun building interleaved problem sets into their materials, and several online learning platforms now offer interleaved practice modes. The evidence base is strong enough that the Institute of Education Sciences in the United States lists interleaving as a recommended practice in its mathematics teaching guides.

The central lesson for any learner is uncomfortable but clear: if practice feels consistently smooth and effortless, it is likely not producing the deep encoding that transfers to novel problems. The difficulty is not a sign that the method is failing. It is the method working.