The effect of deferring feedback on rule-based and information-integration category learning

Previous work has shown that deferring feedback significantly impairs two-dimensional information-integration category learning, often thought to recruit an implicit learning system, but leaves intact unidimensional rule-based learning, commonly assumed to engage an explicit system. These results were taken to support the influential COmpetition between Verbal and Implicit Systems (COVIS) dual-process theory. This conclusion has subsequently been challenged by the finding that this dissociation disappears when the number of relevant dimensions is matched between tasks. However, as well as replacing a unidimensional rule-based task with a two-dimensional conjunction task, a different set of stimuli were used making it unclear which of these alterations was driving the difference in results. The current paper directly examined how both category structure and stimulus type influence the deferred feedback effect. We replicated both the original sets of results but found that deferred feedback also impaired information-integration learning to a greater extent than a conjunction task when the original stimuli were used. These results suggest that the impact of deferred feedback on category learning is more complicated than previously documented, as our findings cannot be easily explained by either COVIS or single-system accounts. Furthermore, our results highlight the critical role that the choice of stimuli has on categorization behavior and emphasize the importance of testing findings across different stimuli to ensure their robustness.