Number and Domain Both Affect the Relation Between Executive Function and Mathematics Achievement: A Study of Children's Executive Function With and Without Numbers

Magnitude processing and executive functions (EFs) have emerged as robust predictors of mathematics achievement. However, the nature of these associations is still unclear. For example, it is uncertain if EFs applied in the context of domain-specific mathematical cognition (i.e., EFs applied while processing numbers) are more closely related to mathematics achievement than EFs applied in nonnumerical, domain-general contexts. Also, how distinct EF domains-that is, working memory, inhibitory control, and cognitive flexibility-and contents-that is, numerical versus nonnumerical-moderate the association between magnitude processing and mathematics achievement has not been fully understood. To address these issues, we investigated how magnitude processing, EFs applied to nonnumerical and numerical task stimuli, and their interactions were associated with mathematics achievement. Three hundred fifty-nine Brazilian third- to fifth-grade (8-10 years old) students completed measures of working memory, inhibitory control, and cognitive flexibility with numerical and nonnumerical task versions, nonsymbolic and symbolic magnitude comparison, and mathematics achievement. A series of regression models indicated that nonsymbolic and symbolic magnitude processing are consistently associated with mathematics achievement, even when controlling for working memory, inhibitory control, and cognitive flexibility measured with both numerical and nonnumerical contents. All EF measures were associated with mathematics achievement. However, cognitive flexibility measured with numerical content showed the strongest association. Results support the hypothesis that magnitude processing and EFs are uniquely associated with mathematics achievement. Furthermore, EFs measured with nonnumerical and numerical contents related differently to mathematics achievement, even when controlling for symbolic and nonsymbolic magnitude processing, suggesting they encompass somewhat distinct cognitive processes. Public Significance Statement<br /> In this study, we showed that nonsymbolic and symbolic magnitude processing and different types of executive functions are predictors of mathematics achievement in a sample of third- to fifth-grade children (8-10 years old) from a developing country. In particular, we observed stronger associations between mathematics achievement and symbolic magnitude processing and cognitive flexibility measured with numerical contents. Interestingly, inhibitory control measured with nonnumerical content had weak correlations with mathematics achievement.