Effects of sample size information and within- and between-category similarity on study choices in self-regulated...

Category learning is usually better supported by interleaved training (alternating between exemplars from different categories) than by blocked training (studying all exemplars within a category sequentially), yet when asked to choose between the two strategies most people endorse blocking as superior. We used a prototype category-learning task to examine the effects of between- and within-category similarity and knowledge of the number of stimuli to be studied on study sequencing choices during self-regulated learning. Across three experiments (including a complete replication), participants who viewed the number of stimuli in each category showed more interleaving in comparison with those who did not, indicating that participants adjusted their strategy based on the projected length of the study phase. Participants informed about the number of stimuli also showed greater interleaving when within-category similarity was high and greater blocking when within-category similarity was low; this difference was not found when participants were not told the number of stimuli to be studied. Between-category similarity did not affect degree of blocking versus interleaving. Overall, interleaving decreased over training and blocking increased. Most participants endorsed hybrid strategies in which blocking was combined with at least some interleaving on a metacognitive questionnaire, but when forced to choose between exclusive blocking and exclusive interleaving, the majority endorsed blocking. These results indicate that participants are sensitive to category structure and expectations about task length when choosing stimuli to study during self-regulated learning, and adjust their strategy across the time course of study.