"Dual-Task Studies of Working Memory and Arithmetic Performance: A Meta-Analysis"

Doctoral student Edward H. Chen (pictured left) is first author and Associate Professor Drew Bailey (pictured right) is co-author of an article in the January issue of the Journal of Experimental Psychology Learning Memory and Cognition.

The article is entitled "Dual-Task Studies of Working Memory and Arithmetic Performance: A Meta-Analysis."

​Chen is a third-year doctoral student specializing in Human Development in Context (HDiC). His research interests include dyscalculia, ADHD, STEM education, memory processes, and out-of-school learning. He is advised by Dr. Bailey.

Bailey researches mathematical development, individual differences, and longitudinal methods.

Abstract

​We performed a meta-analysis of dual-task experiments to assess the robustness of the effects of conducting working memory secondary tasks on arithmetic performance. Four hundred effect sizes from 21 studies from 1,049 participants were analyzed across a variety of specifications. Results revealed that increases in working memory load resulted in slower (7% to 19% reduction) speed of solving of arithmetic problems. Of the potential moderators, working memory load type (i.e. central executive, phonological loop, and visuospatial sketchpad), arithmetic task type (e.g. addition verification, approximate addition, exact multiplication), and authors’ predictions for significance which served as a proxy for cross-talk were statistically significant across specifications, but participants’ age was not. Working memory load type was the most substantial moderator, with central executive tasks leading to the greatest slowing of performance, suggesting that the cognitive complexity of a working memory task may exert a larger influence on performance than the domain-specific overlapping processing demands of similar tasks. We discuss the apparent discrepancy between these findings and findings from correlational studies of the relation between arithmetic performance and working memory, which have reported similar correlations across working memory domains, on average.