Determinants of performance across domains and within bilingualism: Cognitive abilities, experiential factors, and predictability of the task environment.

摘要

How does one become an expert? The present research investigated relationships among general cognitive abilities, experiential factors, and performance variance. An initial study revealed that general cognitive abilities, and not experience, differentiated experts from non-experts in a complex bilingual task. Two longitudinal studies examining performance on the same task followed. The first longitudinal study investigated whether amount of training altered general cognitive abilities. Results indicated that general cognitive abilities recruited during performance in this task improved with experience. The purpose of the second longitudinal study was to investigate 1) individual differences in initial cognitive abilities as predictors of skill acquisition rates, 2) individual differences in cognitive ability change as predictors of skill acquisition rates, and 3) amount of training as a predictor of skill acquisition rates. The following studies were guided by a hypothesis that predictability of the task environment---the rate of change of the task environment while the performer is planning and executing an action, and the range of probable stimuli to which the performer may respond---moderates the relationship between performance variance and its determinants. That is, as predictability of the task environment decreases, the explanatory power of experiential factors decrease and the relative explanatory power of general cognitive abilities increase. For this reason, according to the hypothesis, experiential factors demonstrated a weak direct relationship with performance in the complex bilingual task. In order to test this hypothesis, a laboratory task was designed in which the two features of predictability---rate of change and range of probable stimuli---were manipulated. The results supported the predictability hypothesis: amount of practice was a stronger determinant of performance than general cognitive abilities when the task had low predictability; the opposite pattern emerged when the task had high predictability. Finally, in order to test whether this hypothesis translated into real-world domains, a meta-analysis was conducted examining the relationship between practice and performance variance. The meta-analysis revealed that predictability of the task environment significantly moderated the relationship, such that the lower the predictability, the weaker the relationship between practice and performance. This work advances our understanding of skill acquisition and expertise determinants.