Interleaved Practice Structuring¶

Why It Matters¶

Blocked practice creates an illusion of mastery. When solving 20 similar problems in a row, the solution strategy remains active in working memory - you're not really practicing retrieval or discrimination (knowing which strategy to apply), just executing a cached procedure. This produces smooth performance during practice but failure during tests, where problems are mixed. Interleaving forces you to actively choose the correct strategy for each problem, strengthening both retrieval and discrimination - precisely the skills needed for real-world application.

The Key Move¶

When practicing any skill that involves choosing among multiple approaches, deliberately mix problem types rather than grouping them. After learning multiple techniques/concepts, create practice sets that randomly alternate between them. During practice, you should struggle to identify which approach each problem requires (if it's obvious because you just did three similar problems, you're blocked not interleaved). Accept that practice will feel slower and less smooth - this difficulty is evidence of superior learning.

Classic Application¶

Learning mathematics. Students naturally practice problem sets organized by topic (all quadratic equations together, all word problems together). This feels efficient and produces high success rates during homework. But exam performance is poor because exams mix problem types - students can solve quadratics when told "these are quadratics" but fail to recognize them when mixed with other problems. Interleaved practice (mixed problem sets) forces students to practice discrimination and strategy selection, yielding better exam performance despite worse homework fluency.

Surprising Application¶

Learning to recognize bird species (or any categorization task). Blocked practice: study all sparrow photos together, then all finch photos. You learn features of each category in isolation. Interleaved practice: mix sparrow-finch-warbler photos randomly. This forces you to learn discriminative features (what makes a sparrow distinct from similar birds) rather than just prototypical features (what sparrows look like). Result: dramatically better identification in the field, where birds don't announce their category before appearing.

Failure Modes¶

Excessive difficulty for complete novices - interleaving before any schema is established can be overwhelming; some initial blocked practice may be necessary. Motivation costs - interleaving feels like slower progress, potentially reducing practice volume if learners get discouraged. Domain limitations - purely procedural skills with no discrimination component (typing, playing scales) may not benefit from interleaving. False interleaving - switching between topics without actual mixture (studying chapter 1, then chapter 2, then chapter 3 is not interleaving - it's blocked at chapter level).

Go Deeper¶

Rohrer & Taylor, "The Shuffling of Mathematics Problems Improves Learning" (2007); Kornell & Bjork, "Learning Concepts and Categories: Is Spacing the 'Enemy of Induction'?" (2008); Dunlosky et al., "Improving Students' Learning With Effective Learning Techniques" (2013).