## Definition **Desirable Difficulties** is a concept from cognitive psychology (Robert Bjork, UCLA) describing learning conditions that slow acquisition and reduce short-term performance but produce stronger, more durable, and more flexibly applicable long-term knowledge. The "desirable" qualifier distinguishes productive struggle — which signals importance to memory consolidation — from unproductive cognitive overload, which impedes encoding. Epstein draws on this literature to argue that the most effective learning looks inefficient by conventional metrics. ## Core Mechanisms ### Spacing Distributing practice across time (spaced practice) outperforms massing the same amount of practice in a single session (massed practice), often dramatically. Epstein cites a study in which students who learned Spanish vocabulary and were tested eight years later, having studied the same material, retained 250% more if the original testing had been delayed by a month rather than conducted on the same day. The delay forced deeper encoding. Spacing is counterintuitive because it feels less productive in the moment. ### Interleaving Blocked practice — practising one problem type until fluent, then moving to the next — produces faster short-term gains but weaker long-term retention and transfer. Interleaved practice — mixing problem types within a session — is slower and more frustrating because the learner must first identify the appropriate strategy before applying it, rather than applying a primed strategy automatically. This metacognitive step (identifying what kind of problem this is) is precisely what wicked real-world problems require. A student who only practises "calculate the area of a square" in one block cannot identify which formula applies when a real problem presents ambiguously. ### Testing Effect (Retrieval Practice) Attempting to recall information — even unsuccessfully — strengthens memory more than re-studying the same material. The effort of retrieval, not the success of it, appears to signal memory consolidation. Reading the correct answer after a failed retrieval attempt produces stronger retention than having passively read the answer originally. This is the opposite of how most students study (re-reading, highlighting) and how most teachers teach (demonstrating before testing). ## Procedural vs Connective Questions Epstein distinguishes two question types with divergent effects on understanding: - **Procedural questions** ("What is the formula for the area of a square?") test surface recall and encourage memorisation disconnected from understanding. - **Connective questions** ("Why is area calculated by multiplying two dimensions? What would change if the shape were different?") force the learner to link new knowledge to existing understanding, building the deep, flexible representation that transfers to novel problems. The educational system disproportionately rewards procedural competence; connective understanding is both harder to assess and more valuable. ## The Performance vs Learning Illusion Blocked practice and immediate re-reading produce rapid short-term performance gains that feel like learning but are not durable. Interleaving and spacing feel slower and harder, lowering short-term performance scores — leading teachers and students alike to abandon them in favour of techniques that look productive but are not. Bjork's research repeatedly shows that learners and instructors are poor judges of what techniques produce genuine long-term learning. ## Related - [[Kind vs Wicked Learning Environments]] - [[Generalists vs Specialists]] - [[Late Specialisation and the Sampling Period]] ## Sources - [[Range (Epstein 2019)]]