Relationships between facets of working memory and performance on a curriculum-based mathematics test in children

Abstract

The involvement of different components of working memory in children’s mathematical reasoning was investigated by administering four memory span tasks and a curriculum-based test of mathematics to 9- to 10-year-old children. The mathematics test contained number, space and measurement sections, and the span tasks differed as to whether sequences comprised auditory-verbal or visual-spatial items and whether recall involved the entire sequence (fixed span), or a subset of items from the end of a supraspan sequence (running span). The running span tasks were expected to require updating, and therefore involve executive mechanisms in addition to systems specialised for retaining verbal and spatial information. Surprisingly, the two fixed span variables accounted for more variance than the two running span variables in mathematics test scores. Also, fixed spatial span generally accounted for more variance in mathematics performance than did fixed verbal span, although the latter did make significant unique contributions in predicting scores from two sections of the mathematics test. These outcomes were unchanged when a measure of single word reading was used as a control variable. Discussion centres on the involvement of spatial and verbal short-term memory in different forms of mathematical reasoning. It is argued that running memory tasks may invoke various executive processes depending on their precise structure, and some of these processes may not be critical to mathematical reasoning.