Measuring children's working memory: The influence of titrated time constraints on complex span tasks and the relationship with higher order cognitive abilities

This thesis examined working memory (WM) and high-level cognition (HLC) in children. Previous research has shown that reducing maintenance opportunities in complex span tasks (CSTs) by restricting processing times can strengthen the WM-HLC relationship. This suggests that maintenance strategies are unimportant in the WM-HLC relationship. However, the restriction of processing times equally for all participants has not previously been addressed. This thesis assessed WM in 92 children aged seven to eight years of age using computer-paced numerical, verbal and visuospatial CSTs that titrated processing times individually for each child. Performance was compared to that in a condition where processing times were not restricted. Based on multi-component theories of WM, domain-specific and domain-general relationships with HLC (i.e. nonverbal reasoning, reading, mathematics) were examined. The effects of time constraints on the underlying mechanisms of each CST (storage, processing time, recall time, processing accuracy), their relationships with each other, and with HLC were investigated. In addition, the contributions of the broader executive abilities of inhibition and task-switching to the WM-HLC relationship were examined. Finally, the link between current WM abilities and mathematics performance two years later was also explored. Results showed that the two administration conditions accounted for shared and unique variance in HLC, suggesting measurement of different and similar cognitive abilities important in certain higher-order cognitive tasks. Examination of the underlying CST mechanisms showed that numerical WM best predicted concurrent HLC, with processing time replacing storage as a predictor when time constraints were introduced. Longitudinally, numerical, verbal and visuospatial WM predicted mathematics two years later. This identified WM capacity in seven to eight year olds important in mathematical ability at the ages of nine to ten years. Task-switching and inhibition did not predict HLC. Implications for multi-component and attention-based theories of WM, the importance of processing speeds and the role of maintenance strategy in the WM-HLC relationship are discussed.