The starting point for understanding effective learning is Hermann Ebbinghaus's forgetting curve: memory decays exponentially after initial encoding, with roughly half of new material lost within an hour and up to seventy percent gone within twenty-four hours without reinforcement. Spaced repetition counteracts this decay by timing reviews just before forgetting would occur, with each successful retrieval resetting the memory trace and making the forgetting curve progressively shallower. Distributing study across multiple sessions (distributed practice) consistently outperforms cramming the same hours into one sitting (massed practice), even though massed practice produces a stronger short-term sense of mastery. Over the long term, frequent short exposures to material build far more durable memory than marathon study sessions.
Retrieval practice, also called the testing effect, is the single most powerful learning technique identified in laboratory research. The act of pulling information out of memory, rather than passively re-reading it, strengthens the memory trace far more than equivalent time spent reviewing, with studies showing that a single short quiz after a lecture roughly doubles retention after one week compared to re-reading the same material. Retrieval practice is not just an assessment; it is the primary learning event. When combined with spaced repetition, each timed retrieval both consolidates the memory and signals to the scheduling system when to bring the material back.
The generation effect extends this principle: actively producing an answer, even an incorrect one, leads to better memory than simply reading the answer. The mental effort of generation creates richer, more distinctive memory traces because the learner must engage prior knowledge to produce a response. Generation is one example of a broader principle called desirable difficulty: learning conditions that feel harder in the moment, including retrieval practice, spacing, interleaving, and errorful learning, almost always produce stronger long-term retention than easier study methods, even though they feel less productive while you are doing them.
Several frameworks organize these findings. The AGES model highlights four levers: Attention (focused engagement), Generation (creating your own connections), Emotion (emotional relevance), and Spacing (distributed practice). The principle of transfer-appropriate processing reminds us that memory is best retrieved when the conditions at recall match those at encoding, while encoding specificity emphasizes that retrieval cues work best when they match the original learning context. Because declarative memory (facts and events) and procedural memory (skills and habits) rely on different brain systems, learning strategies must be matched to the type of material being learned.