In one sentence
Mixing categories is not automatically best: this study suggests interleaving fits learning by resemblance, while blocking may better support finding a rule.
What the researchers did
Little and Nepangue ran experiments on category learning that replicated an earlier comparison of study order and learning strategy. Participants learned categories either by relying on remembered similarity between examples or by trying to discover a rule that defined the category.
Examples were presented either interleaved, with categories alternating, or blocked, with examples from one category grouped together. The researchers tested transfer immediately after learning and again after roughly 48 hours. They asked whether the match between the study order and the learner’s strategy would last beyond the first test.
What they found
- For learning based on similarity and memory, interleaving was more effective than blocking.
- For learning based on a rule, blocking was more effective than interleaving.
- This interaction between strategy and study order appeared again on the delayed test about 48 hours later.
- The results also suggested that some rule-based learning was more resistant to forgetting than memory-based learning.
- There is no single presentation order that won for every learning goal.
What this means for learners and educators
- Mix problem types when the task is to notice distinctions among similar cases—for example, separating closely related concepts or visual categories.
- Group examples when learners first need to infer a stable rule, such as a grammatical pattern or a mathematical procedure.
- Before changing a study schedule, name the goal: “Am I memorizing examples, comparing cases, or extracting a rule?” That answer can guide the order.
Limitations and what we don't know yet
The abstract does not provide participant numbers or detailed materials, so it cannot show how large the effects were or who was represented. The delay was about two days, not a school term or a final exam period. “Memory-based” and “rule-based” learning may also overlap in real classrooms, where a learner can switch strategies. The study supports matching sequence to a goal; it does not supply a universal schedule for all subjects.