A review of the cultural psychology literature reveals that some ethnic groups consistently perform differently on even the most basic cognitive tasks. Specifically, Asians attend to more contextual information whereas Caucasians selectively attend to the most salient stimulus dimension. In order to determine if such processing differences in attention impact category learning, this dissertation investigated whether Chinese, Caucasians, and Latinos performed differently on the perceptual categorization task.

Seventy-two Caucasian, 50 Chinese, and 47 Latino students matched in terms of years of education and gender represented a range of acculturation from foreign born bilingual immigrants to native born Americans whose only language is English. Participants learned to sort stimuli into one of two predetermined categories by receiving corrective feedback after each trial. In Experiment 1, participants learned a unidimensional rule requiring participants to attend selectively to a single stimulus dimension. In Experiment 2, participants learned a conjunctive rule requiring an explicit conjunction of two stimulus dimensions. In Experiment 3, participants learned an information-integration rule requiring that participants integrate information from two stimulus dimensions at an implicit level. In addition to examining potential differences in accuracy using repeated measures analysis of variance, mathematical models identified the types of categorization strategies participants actually used when learning one of the three categorization rules. Finally, a regression analysis further investigated the possible underpinnings of observed ethnic group differences in categorization accuracy.

No differences were observed between groups in Experiments 1 and 3, suggesting that category learning tasks that emphasize selective attention and implicit learning processes are not sensitive to the ethnic differences observed in previous studies. In Experiment 2, however, contrary to what would have been expected, Caucasians performed better than both Chinese and Latino participants when having to form an explicit sorting rule combining more than one stimulus dimension. Even after groups were equated based on the type of categorization strategies employed, these group differences persisted. Hierarchical regression further revealed that ethnicity did not predict accuracy after level of mainstream acculturation was taken into account. Secondary analysis of the acculturation subscales suggested that receptive language ability in English was the best predictor of overall accuracy in learning a conjunctive rule-based task over and above ethnic group membership. Overall, these results suggest that previously observed cultural differences are more likely related to factors other than ethnic group membership, and therefore warrant further study.